Integrating variant extensions from multiple sequencing tubes

Abstract

The present disclosure describes methods, non-transitory computer-readable media, and systems that can generate genotype detections from a combined pipeline used to process nucleotide reads from multiple read types / sources to achieve robust, accurate genotype detections. For example, the disclosed systems may train and / or utilize a genotype detection integrated machine learning model to generate predictions for genotype detection based on data associated with a first type of nucleotide reads (e.g., short reads) and a second type of nucleotide reads (e.g., long reads). As disclosed, the disclosed systems may determine sequencing metrics and may utilize a genotype detection integrated machine learning model to generate predictions for generating output genotypes based on the sequencing metrics (e.g., genotype probabilities, variant detection classifications). The disclosed system can utilize a plurality of such genotype detection integrated machine learning models to generate genotype detections for different variant types, such as SNPs and insertion deletions, wherein the genotype detection integrated machine learning models generate different predictions for each variant type.

Description

[0001] CROSS-REFERENCE TO RELATED APPLICATIONS

[0002] This application claims the benefit of and priority to U.S. Provisional Application No. 63 / 482,163, filed on January 30, 2023, entitled “INTEGRATING VARIANT CALLS FROM MULTIPLE SEQUENCING PIPELINES UTILIZING A MACHINE LEARNING ARCHITECTURE” and U.S. Provisional Application No. 63 / 378,474, filed on October 5, 2022, entitled “INTEGRATING VARIANT CALLS FROM MULTIPLE SEQUENCING PIPELINES UTILIZING A MACHINE LEARNING ARCHITECTURE”. The above applications are hereby incorporated by reference in their entirety. Background Art

[0003] In recent years, biotechnology companies and research institutions have improved the hardware and software for sequencing nucleotides and determining the nucleotide base calls of reads, and subsequently determining the variant calls and genotype calls of genomic samples. For example, some existing nuclear base sequencing platforms determine the individual nucleotide bases (or "nucleobases") within a sequence by using conventional Sanger sequencing or by using a sequencing-by-synthesis (SBS) method. When using SBS, existing platforms can monitor thousands of nucleic acid polymers synthesized in parallel to predict genotype calls based on larger base call data sets. For example, cameras in many SBS platforms capture images of irradiated fluorescent labels incorporated into oligonucleotides for determining nuclear base calls. After capturing such images, existing SBS platforms send base call data (or image data) to a computing device to apply sequencing data analysis software that determines the nuclear base sequence of nucleic acid polymers. Based on the differences between the aligned nucleotide reads and the reference genome, existing systems can further utilize variant callers to identify variants of genomic samples, such as single nucleotide polymorphisms (SNPs), insertions and deletions (indels) and / or structural variants, as well as genotype calls.

[0004] Despite these recent advances in sequencing and variant detection, existing sequencing systems typically include variant detectors that cannot accurately determine variant detection, particularly for SNPs and insertions and deletions. For example, many existing systems generate variant detections that include excessive false positive detections and / or false negative detections for SNPs and insertions and deletions. The reason for this inaccuracy is that the constraints of some existing sequencing systems dictate that they generate variant detections from a single-stream processing pipeline that focuses on only one read source at a time. For example, as mentioned above, some existing systems perform variant detection and / or variant detection filtering based only on nucleotide reads from SBS sequencing. For another example, some existing systems perform variant detection based only on nucleotide reads from certain types of long reads such as circular consensus sequencing (CCS) reads or nanopore long reads. Therefore, reliance on a single source of read data causes many existing systems to generate variant calls that include excessive false positive calls and / or false negative calls for certain clinical benchmarks, which could be reduced by more accurate systems. Further complicating the issue is that different sequencing systems exhibit different error distributions, such as when existing systems based on CCS reads and nanopore long reads generate variant calls with higher indel errors relative to sequencing systems using other types of reads.

[0005] Compounding this inaccuracy in variant calling is that some existing sequencing systems utilize models that require training on millions or billions of base call data that is either unavailable or incomplete. More specifically, some existing sequencing systems utilize deep learning models that require large amounts of training data to achieve acceptably accurate measurements. However, for certain variant types (e.g., structural variants), the training data for the variants is relatively limited, and training models using incomplete or insufficient data can result in inaccurate and unreliable variant calling predictions. As a result, some existing systems that rely on deep learning models may produce inaccurate variant calls, including SNPs and indels.

[0006] In addition to inaccurately determining variant calls, some existing sequencing systems also consume computing resources inefficiently due to overly complex models. Specifically, the variant callers of some existing sequencing systems are computationally expensive and slow. In fact, some existing sequencing systems utilize variant callers with deep learning architectures, which require a large amount of computing resources (e.g., computing time, processing power, and memory) to train and apply deep learning architectures. For example, some existing sequencing systems require hundreds of hours and multiple graphics processing units (GPUs) to train complex convolutional neural networks or other deep learning architectures, which require many hours (e.g., up to 24 hours) across multiple computing devices to generate variant calls or genotype calls for a single sample sequence even after training.

[0007] As another shortcoming of existing sequencing systems with complex deep learning networks, many such systems utilize model architectures that render sequence data uninterpretable. More specifically, as a basis for generating variant calls, some existing deep neural networks transform and manipulate sequence data multiple times, transforming from one uninterpretable latent vector to another such latent vector across various layers and neurons during processing. In many cases, the internal data of these deep neural networks is uninterpretable and difficult to exploit in any way outside of the neural network architecture itself. Summary of the invention

[0008] The present disclosure describes embodiments of methods, non-transitory computer-readable media, and systems that can utilize machine learning models to generate predictions for genotype detection based on data from different types of nucleotide reads. Specifically, the disclosed system can generate genotype calls from a combined pipeline for processing nucleotide reads from multiple read types / sources to achieve robust, accurate genotype calls (including compositional variant calls). For example, the disclosed system can train or utilize a genotype call integrated machine learning model to generate predictions for genotype calls based on data associated with a first type of nucleotide read (e.g., short read) and a second type of nucleotide read (e.g., long read). As disclosed, the system can determine sequencing metrics for a first genotype call corresponding to a first type of nucleotide read and a second genotype call corresponding to a second type of nucleotide read. Based on different or shared sequencing metrics corresponding to the first genotype call and the second genotype call, the disclosed system utilizes a genotype call integrated machine learning model to generate predictions (e.g., genotype probabilities, variant call classifications) for updating or confirming a first genotype call or a second genotype call, or determining different genotype calls. In some cases, the disclosed system may utilize multiple such genotype call ensemble machine learning models to update or confirm genotype calls for different variant types (such as SNPs and indels), where the genotype call ensemble machine learning models generate different predictions for each variant type. BRIEF DESCRIPTION OF THE DRAWINGS

[0009] The detailed description refers to the accompanying drawings which are briefly described below.

[0010] Figure 1 A block diagram of an example computing environment for implementing a sequencing system and an integrated detection system according to one or more embodiments is illustrated.

[0011] Figure 2 An overview of a call integration system for generating genotype calls using a genotype call integration machine learning model according to one or more embodiments is illustrated.

[0012] Figure 3 Illustrated are example types of nucleotide reads based on which a call integration system can generate genotype calls according to one or more embodiments.

[0013] FIG. 4A to FIG. 4C An integrated system for determining sequencing metrics that are shared or different between different types of nucleotide reads according to one or more embodiments is illustrated.

[0014] FIG. 5A to FIG. 5CIllustrated is a calling integration system that utilizes a genotype calling integration machine learning model to generate predictions (e.g., genotype probabilities or variant call classifications) and corresponding genotype calls according to one or more embodiments.

[0015] Figure 6 An example diagram illustrating a training process for learning parameters of a genotype calling ensemble machine learning model according to one or more embodiments.

[0016] Figure 7 An example diagram is illustrated for updating or generating a merged variant call file based on predictions of a genotype calling ensemble machine learning model according to one or more embodiments.

[0017] Figure 8 Example graphs and tables illustrating accuracy metrics for a detection integration system according to one or more embodiments.

[0018] Fig. 9A An example table illustrating accuracy metrics for a detection integration system according to one or more embodiments is shown.

[0019] Fig. 9B An example table illustrating accuracy metrics for a detection integration system according to one or more embodiments is shown.

[0020] FIG. 10A to FIG. 10B A graph depicting accuracy metrics associated with a detection integration system is illustrated according to one or more embodiments.

[0021] Fig.11 A flow chart illustrating a series of actions for utilizing a genotype call ensemble machine learning model to generate genotype calls based on nucleotide reads of a first read type and a second read type, according to one or more embodiments.

[0022] Fig.12 A block diagram of an example computing device for implementing one or more embodiments of the present disclosure is illustrated. DETAILED DESCRIPTION

[0023] The present disclosure describes an embodiment of a detection integration system that generates and modifies genotype detection for genomic samples using a genotype detection integration machine learning model. Specifically, the detection integration system can generate output genotype detection (e.g., reported genotype detection from a merged variant detection file) based on multiple initial genotype detections (e.g., variant detections) using a genotype detection integration machine learning model, and the output genotype detection is for a genomic locus generated by a detection generation model according to different read types. In order to generate output genotype detection, in certain embodiments, the detection integration system generates or receives initial genotype detection from read data associated with a combination of short reads (e.g., sequencing by synthesis or "SBS" reads) and long reads (e.g., nanopore long reads, cycle common sequencing or "CCS" reads and / or assembled nucleotide reads). In some cases, the detection integration system determines or identifies a specific sequencing metric (e.g., based on read data, detection generation model data and / or external data) to be input into a genotype detection integration machine learning model for generating output genotype detection. The call integration system can further train or apply a genotype call integration machine learning model based on the sequencing metrics to generate (or refine or recalibrate) genotype calls.

[0024] As just mentioned, in certain specific implementations, the detection integration system uses read data from different read types to improve genotype detection accuracy (and corresponding variant detection accuracy). In order to facilitate the generation of genotype calls based on multiple read types, in some embodiments, the detection integration system receives initial genotype calls from the detection generation model. For example, the detection integration system (i) receives or determines an initial genotype call corresponding to a first type of nucleotide read (e.g., a short read) (e.g., a call indicating a genotype at the genomic coordinates of a nucleotide sequence), and further (ii) receives or determines another initial genotype call corresponding to a second type of nucleotide read (e.g., a long read). In some cases, the first type of nucleotide read includes nucleotide reads synthesized from sample library fragments that are shorter than a first threshold number of nucleobases. Conversely, in the same or other circumstances, the second type of nucleotide reads includes (i) assembled nucleotide reads that have been assembled from shorter nucleotide reads to form a contiguous sequence that satisfies a second threshold number of nucleobases, (ii) CCS reads that satisfy the second threshold number of nucleobases, and / or (iii) nanopore long reads that satisfy the second threshold number of nucleobases.

[0025] According to the initial genotype detection corresponding to different read types, the detection integration system can further generate output genotype detection, such as prediction of the presence or absence of variants (such as SNPs or insertions and deletions) and the zygosity of alleles of genomic samples. As mentioned, in order to generate output genotype detection, the detection integration system can extract, identify or determine sequencing metrics (associated with initial genotype detection from different read types) to input into the genotype detection integration machine learning model. Then, the genotype detection integration machine learning model generates a possibility or prediction set indicating the possibility that the initial genotype detection is correct or incorrect (e.g., for each initial genotype detection corresponding to different read types and / or for each different variant type Different prediction sets). For example, the detection integration system can extract or determine sequencing metrics belonging to one or more categories, including: (i) sequencing metrics based on reads, (ii) sequencing metrics generated by the detection model, and (iii) sequencing metrics from external sources. Additional details on the composition and determination of sequencing metrics are provided below with reference to the accompanying drawings.

[0026] As proposed, in certain embodiments, the call integration system generates genotype calls using a multi-stream pipeline that processes multiple read types of output genotype calls based on multiple read types as part of a combined or merged variant call file. For example, the call integration system (i) processes a first set of sequencing metrics extracted from an initial genotype call based on a first read type, and (ii) processes a second set of sequencing metrics extracted from an initial genotype call based on a second read type. In addition, the call integration system can utilize a genotype call integration machine learning model to generate a prediction set based on the first set of sequencing metrics and the second set of sequencing metrics, and can generate an output genotype call based on the prediction set.

[0027] In order to predict or generate output genotype calls for different variant types (e.g., SNPs and indels), in some cases, the call integration system generates a set of different predictions for different variant types using a genotype call integration machine learning model (e.g., based on the same or different sequencing metrics). For example, the call integration system may utilize a first instance of a genotype call integration machine learning model (e.g., trained to predict SNPs) to process a first read type sequencing metric (e.g., an SBS sequencing metric) and a second read type sequencing metric (e.g., an assembled nucleotide read sequencing metric) to generate an output genotype call for a SNP at a genomic coordinate. In addition, the call integration system may utilize a second genotype call integration machine learning model (e.g., trained to predict indels) to process a first read type sequencing metric and a second read type sequencing metric to generate an output genotype call for an indel at different (or the same) genomic coordinates. In some embodiments, the detection integration system may utilize a first genotype detection integration machine learning model for biallelic SNPs, and may utilize a second genotype detection integration machine learning model for other types of variant detection (e.g., variants that are not biallelic SNPs). In addition, although the present disclosure describes at least two different types of genotype detection integration machine learning models, in certain specific implementations, the detection integration system trains or applies a single genotype detection integration machine learning model to generate two genotype predictions for different types of variants (e.g., genotype predictions for SNPs or indels).

[0028] As mentioned above, the detection integration system provides several advantages, benefits and / or improvements over existing sequencing systems (including variant detectors and other sequencing data analysis software). For example, the detection integration system generates more accurate genotype calls (including variant calls) than existing sequencing systems. Although some existing sequencing systems do not accurately generate variant calls (especially for SNPs and insertions and deletions), the detection integration system trains or uses genotype detection integration machine learning models to improve genotype / variant calls that are superior to existing systems. Specifically, unlike existing systems that rely on a single source of read data, the detection integration system can process multiple reads of different types (e.g., assembled nucleotide reads and SBS reads) to generate more accurate genotype calls corresponding to SNPs and insertions and deletions (thereby reducing false positives and false negatives). In addition, the reason for the accuracy being superior to existing systems is that the detection integration system can use different instances of genotype detection integration machine learning models trained for different variant types (e.g., SNPs and insertions and deletions) to generate or predict genotype calls based on multiple read types, which is not possible with existing systems. Further, the improved accuracy of genotype calls is due to the fact that, in some cases, the call integration system determines and utilizes specific sequencing metrics (which are different from existing systems) as the basis for generating calls (e.g., as input data) via the genotype call integration machine learning model.

[0029] In order to achieve the above-mentioned improved accuracy, the call integration system utilizes an improved and unique machine learning model that is trained to perform new applications, namely a genotype call integration machine learning model. Unlike existing variant calls that generate genotype calls based on general single-stream sequencing data without adjusting or emphasizing whether specific genomic coordinates have historically exhibited or have been detected to exhibit specific variants, the call integration system utilizes a unique genotype call integration machine learning model (multiple instances) that generates specific predictions or classifications for different types of variants (e.g., SNPs and insertions and deletions) based on multi-read type data. In some cases, the call integration system utilizes a genotype call integration machine learning model as a post-processing filter to (i) select between a first genotype call corresponding to a first type of nucleotide read and a second genotype call corresponding to a second type of nucleotide read, or (ii) determine another genotype call that is different from the first genotype call and the second genotype call.

[0030] The reason for the improvement in accuracy, at least in part, is that the detection integration system exhibits improved flexibility over existing sequencing systems. For example, while many existing sequencing systems are limited to analyzing read data from one read type at a time, in some embodiments, the detection integration system is suitable for processing multiple read types to merge data and generate output genotype calls for specific genomic coordinates or regions. Specifically, unlike some existing sequencing systems, the detection integration system can generate genotype calls (e.g., including variant calls) for genomic coordinates based on multiple types of read data (such as assembled nucleotide reads and SBS reads) of genomic coordinates.

[0031] In addition to improved accuracy and flexibility, in certain embodiments, the detection integrated system also improves computational efficiency and speed. As described above, some existing sequencing systems utilize computationally expensive, slow neural network architectures (e.g., deep learning architectures, such as convolutional neural networks), which require several hours (e.g., up to 24 hours) across multiple high-end processors to achieve the detection for processing read segment data to generate detection for genomic samples. In addition, the detection integrated system can generate (merged) variant detection files by only updating certain fields, without regenerating brand-new variant detection files (as some existing systems do). This deep learning architecture may also require several days (or weeks) to train. On the contrary, the detection integrated system utilizes a relatively lightweight, fast architecture for genotype detection integrated machine learning models. Compared with the need to spend several hours across multiple processors for existing sequencing systems, the detection integrated system requires less than an hour (e.g., about fifteen minutes for detection generation models, less than one minute for genotype detection integrated machine learning models) of running time (e.g., on a single processor) to generate genotype detection (and / or variant detection) of genomic samples. Therefore, the Call Ensemble system is much faster and computationally cheaper than many deep learning methods for genotype / variant calling. In fact, not only are the Call Ensemble system's models faster and computationally cheaper to implement, but the Genotype Call Ensemble machine learning models are also much faster and computationally cheaper than many existing deep learning systems.

[0032] As an additional advantage over existing sequencing systems, in certain specific implementations, the detection integration system can identify or promote changes in various sequencing metrics that affect the accuracy of genotype calls (and corresponding variant calls). Although the neural network architecture of many existing sequencing systems makes it impossible to use the potential features hidden between its many layers and neurons to interpret the internal model data, the detection integration system utilizes a model architecture that facilitates the interpretation of the effects of various sequencing metrics. More specifically, in some cases, the detection integration system utilizes a detection generation model and a detection integration machine learning model that makes it easier to extract and analyze the various sequencing metrics used throughout the process of generating genotype calls. In fact, the detection integration system can determine the corresponding contribution measure of the sequencing metrics involved in the genotype call at a specific region of a specific genomic coordinate.

[0033] As proposed in the foregoing discussion, the present disclosure utilizes a variety of terms to describe the features and advantages of the detection integrated system. Additional details about the meaning of these terms used in the present disclosure are provided below. As used in the present disclosure, for example, the term "sample nucleotide sequence" or "sample sequence" refers to a nucleotide sequence (or a copy of such a separated or extracted sequence) separated or extracted from a sample organism. Specifically, the sample nucleotide sequence includes a fragment of a nucleic acid polymer that is separated or extracted from a sample organism and is composed of nitrogen-containing heterocyclic bases. For example, the sample nucleotide sequence may include a fragment or molecule of deoxyribonucleic acid (DNA), ribonucleic acid (RNA), or other polymeric forms of nucleic acids or chimeric or hybrid forms of nucleic acids described below. More specifically, in some cases, the sample nucleotide sequence is present in a sample prepared or separated by a test kit and received by a sequencing device.

[0034] Relatedly, as used herein, the term "genomic sample" refers to a target genome or a part of a genome that is subjected to determination or sequencing. For example, a genomic sample includes one or more nucleotide sequences (or a copy of such a sequence separated or extracted) separated or extracted from a sample organism. Specifically, a genomic sample includes a full genome separated or extracted (in whole or in part) from a sample organism and composed of nitrogen-containing heterocyclic bases. A genomic sample may include a fragment or molecule of deoxyribonucleic acid (DNA), ribonucleic acid (RNA), or other polymeric forms of nucleic acid or a chimeric or heterozygous form of nucleic acid described below. In some cases, a genomic sample is present in a sample prepared or separated by a test kit and received by a sequencing device.

[0035] As further used herein, the term "genotype detection" refers to determining or predicting the specific genotype of a genomic sample at a genomic locus. Specifically, genotype detection may include predicting the specific genotype of a genomic sample relative to a reference genome or reference sequence at a genomic coordinate or genomic region. For example, in some cases, genotype detection includes determining or predicting that a genomic sample includes both a nuclear base and a complementary nuclear base at a genomic coordinate, and the genomic coordinate is homozygous or heterozygous for a reference base or variant (for example, a homozygous reference base is expressed as 0|0, or heterozygous for a variant on a specific chain, expressed as 0|1). Therefore, genotype detection may include predicting variants or reference bases of one or more alleles of a genomic sample, and indicating the zygosity of a variant or a reference base. Genotype detection is typically determined for genomic coordinates or genomic regions, at which SNPs, insertions, deletions or other variants have been identified for an organism population.

[0036] In some cases, an "initial genotype call" refers to a genotype call corresponding to or determined based on nucleotide read data and / or sequencing metrics for a particular type of nucleotide read. For example, an initial genotype call may include a first genotype call corresponding to a first type of nucleotide read of a first threshold number of nucleobases and / or a second genotype call corresponding to a second type of nucleotide read of a second threshold number of nucleobases. In contrast, an "output genotype call" refers to a genotype call reported by an output data file or generated for an output data file. For example, an output genotype call includes a final genotype call determined based on one or both of a genotype probability and a variant call classification from a genotype call ensemble machine learning model and included in a variant call file (VCF).

[0037] As further used herein, the term "nucleobase call" (or simply "base call") refers to the determination or prediction of a specific nucleobase (or nucleobase pair) of a genomic coordinate of an oligonucleotide (e.g., a nucleotide read) or a sample genome during a sequencing cycle. Specifically, a nucleobase call can indicate: (i) the determination or prediction of the type of nucleobase that has been incorporated into an oligonucleotide on a nucleotide sample slide (e.g., a read-based nucleobase call); or (ii) the determination or prediction of the type of nucleobase present at a genomic coordinate or region within a genome, including variant calls or non-variant calls in a digital output file. In some cases, for a nucleotide read, a nucleobase call includes determining or predicting a nucleobase based on an intensity value generated by a fluorescently labeled nucleotide of an oligonucleotide added to a nucleotide sample slide (e.g., in a cluster of a flow cell). Alternatively, a nucleobase call includes determining or predicting a nucleobase based on chromatographic peaks or current changes generated by nucleotides passing through a nanopore of a nucleotide sample slide. In contrast, based on the nucleotide reads corresponding to the genomic coordinates, the nucleobase call may also include the nucleobase at the genomic coordinates of the sample genome of the final predicted variant call file (VCF) or another base call output file. Therefore, the nucleobase call may include base calls corresponding to the genomic coordinates and the reference genome, such as an indication of a variant or non-variant at a specific position corresponding to the reference genome. In fact, the nucleobase call may refer to a variant call, including but not limited to a single nucleotide variant (SNV), an insertion or deletion (indel), or a base call as part of a structural variant. As mentioned above, a single nucleobase call may be an adenine (A) call, a cytosine (C) call, a guanine (G) call, a thymine (T) call, or a uracil (U) call.

[0038] Relatedly, as used herein, the term "nucleotide read" refers to a sequence of one or more nucleotide bases (or nucleobase pairs) inferred from all or part of a sample nucleotide sequence (e.g., a sample genomic sequence, complementary DNA). Specifically, a nucleotide read includes a sequence determined or predicted from a sequencing library corresponding to a genomic sample based on the nucleobase detection of a nucleotide fragment (or a monoclonal nucleotide fragment group). For example, in some embodiments, the detection integrated system determines the nucleotide read by generating a nucleobase detection of a nucleobase passing through a nanopore of a nucleotide sample slide, determined via fluorescent labeling, or determined based on a hole in a circulation pool. In some cases, a nucleotide read may refer to a specific type of read, such as a nucleotide read synthesized from a sample library fragment shorter than a threshold number of nucleobases (e.g., an SBS read). In these or other cases, another type of nucleotide read may refer to (i) an assembled nucleotide read that has been assembled from shorter nucleotide reads to form a contiguous sequence that satisfies a threshold number of nucleobases (e.g., an assembled nucleotide read), (ii) a cycle consensus sequencing (CCS) read that satisfies a threshold number of nucleobases, or (iii) a nanopore long read that satisfies a threshold number of nucleobases.

[0039] As described above, in some embodiments, the detection integration system determines the sequencing metrics of the nucleobase calls of the nucleotide reads. As used herein, the term "sequencing metric" refers to a quantitative measurement or score that indicates the degree to which a single nucleobase call (or sequence of a nucleobase call) is aligned, compared, or quantified relative to the genomic coordinates or genomic regions of a reference genome, relative to nucleobase calls from nucleotide reads, or relative to external genome sequencing or genomic structure. For example, sequencing metrics include quantitative measurements or scores indicating the following degrees: (i) a single nucleobase call is aligned, mapped, or covers the genomic coordinates or reference bases of a reference genome; (ii) a nucleobase call is compared with a reference or alternative nucleotide read in terms of mapping, mismatch, base call quality, or other raw sequencing metrics; or (iii) a genomic coordinate or region corresponding to a nucleobase call exhibits mappability, repetitive base call content, DNA structure, or other broad metrics.

[0040] Along these lines, the detection integration system determines various types of sequencing metrics from different sources, such as sequencing metrics based on reads, sequencing metrics from external sources, and sequencing metrics generated by detection models. As used herein, the term "sequencing metrics based on reads" refers to sequencing metrics of nucleotide reads derived from sample nucleotide sequences. For example, sequencing metrics based on reads include sequencing metrics determined by applying statistical tests to detect differences between reference sequences and nucleotide reads. In some embodiments, sequencing metrics based on reads may include comparative mapping quality distribution metrics indicating comparisons between mapping qualities or comparative mismatch count metrics indicating comparisons between mismatch counts. In some cases, sequencing metrics based on reads may correspond to nuclear base calls generated according to different read types (such as assembled nucleotide reads and / or SBS reads).

[0041] In contrast, "sequencing metrics from an external source" refers to sequencing metrics identified or obtained from one or more external databases. For example, sequencing metrics from an external source include metrics related to nucleotide mappability, replication timing, or DNA structure that can be obtained outside of the detection integration system.

[0042] In addition, the term "sequencing metrics generated by the detection model" refers to internal, model-specific sequencing metrics generated or extracted by the detection generation model. For example, the sequencing metrics generated by the detection model include variant detection sequencing metrics extracted or determined via the variant detector component of the detection generation model and mapping and alignment sequencing metrics extracted or determined via the mapping and alignment component of the detection generation model. As indicated above, the sequencing metrics generated by the detection model may include alignment metrics that quantify the degree to which the sample nucleic acid sequence is aligned with the genomic coordinates of the example nucleic acid sequence, such as a deletion size metric or a mapping quality metric. In addition, the sequencing metrics generated by the detection model may include depth metrics that quantify the depth of the nucleobase detection of the sample nucleic acid sequence at the genomic coordinates of the example nucleic acid sequence, such as a forward-reverse depth metric or a normalized depth metric. The sequencing metrics generated by the detection model may also include detection quality metrics that quantify the quality or accuracy of the nucleobase detection, such as a nucleobase detection quality metric, a detectability metric, or a somatic cell quality metric.

[0043] As further used herein, the term "genomic coordinates" (or sometimes simply "coordinates") refers to a specific position or orientation of a nucleobase within a genome (e.g., a genome of an organism or a reference genome). In some cases, the genomic coordinates include an identifier for a specific chromosome of the genome and an identifier for the position of the nucleobase within a specific chromosome. For example, one or more genomic coordinates may include a chromosome number, name, or other identifier (e.g., chr1 or chrX) and one or more specific positions, such as a numbered position after the chromosome identifier (e.g., chr1:1234570 or chr1:1234570-1234870). In some cases, genomic coordinates refer to genomic coordinates on sex chromosomes (e.g., chrX or chrY). Therefore, the detection integration system can determine the genotype probability and / or variant detection classification of a genotype detection (e.g., variant detection) for a genomic coordinate on a sex chromosome. In addition, in certain embodiments, the genomic coordinates refer to the source of the reference genome (e.g., mt for a mitochondrial DNA reference genome or SARS-CoV-2 for a reference genome of the SARS-CoV-2 virus) and the position of the base in the source of the reference genome (e.g., mt:16568 or SARS-CoV-2:29001). In contrast, in some cases, the genomic coordinates refer to the position of the base in the reference genome without reference to the chromosome or source (e.g., 29727).

[0044] In addition, as used herein, the term "genomic region" refers to a range of genomic coordinates. Like genomic coordinates, in certain embodiments, a genomic region can be identified by a chromosome identifier and one or more specific positions, such as the numbered positions after the chromosome identifier (e.g., chr1: 1234570-1234870).

[0045] As mentioned above, genome coordinates include positions within a reference genome. Such positions may be within a specific reference genome. As used herein, the term "reference genome" refers to a digital nucleic acid sequence assembled as a representative example (or multiple representative examples) of genes and other genetic sequences of an organism. Regardless of sequence length, in some cases, a reference genome represents an exemplary set of genes or a set of nucleic acid sequences determined by scientists to represent a specific species in a digital nucleic acid sequence of an organism. For example, a linear human reference genome may be GRCh38 or other versions of a reference genome from a genome reference consortium. As another example, a reference genome may include a reference map genome containing a linear reference genome and a path representing a nucleic acid sequence from an ancestral haplotype, such as an Illumina DRAGEN map reference genome hg19.

[0046] Additionally, as used herein, the term "reference multigenome" (sometimes referred to as a "map reference genome") refers to a reference genome that includes both a linear reference genome and an alternative continuous sequence (or map amplification) representing a variant haplotype sequence or other variant or alternative nucleic acid sequence. For example, a reference multigenome may include a linear reference genome and an alternative continuous sequence corresponding to one or more population haplotype sequences identified from a genomic sample database. As an example, a reference multigenome may include the Illumina DRAGEN map reference genome hg19.

[0047] As further used herein, the term "continuous sequence" (or "continuous assembly") refers to a common nucleotide sequence of a genomic region of a genomic sample (or multiple genomic samples of a species), which is based on a set of overlapping nucleotide segments corresponding to the genomic region. Specifically, a continuous sequence includes a common nucleotide sequence of a genomic region of one or more genomic samples, which is based on nucleotide reads of one or more genomic samples covering (or overlapping) the genomic region. As described above, the terms "continuous sequence" and "continuous assembly" are used interchangeably.

[0048] Relatedly, the term "alternative contiguous sequence" (or simply "alternative contiguous") refers to a contiguous sequence representing a population haplotype added to a linear reference genome (or other reference genome) at one or more specific genomic coordinates (e.g., promoted to a linear reference genome). In some embodiments, a map reference genome (or reference multi-genome) may include an alternative contiguous sequence that maps to the genomic coordinates of a primary assembly of a linear reference genome. For example, an alternative contiguous sequence may represent a population haplotype comprising a variant having a promotion of two or more genomic coordinates in a linear reference genome corresponding to two or more flanks of a variant breakpoint. In some cases, a hash table of a map reference genome (or reference multi-genome) includes an identifier that associates an alternative contiguous sequence representing a variant haplotype with a genomic coordinate representing a reference haplotype from a primary assembly of a linear reference genome.

[0049] As used herein, the term "base call quality metric" refers to a specific score or other measure that indicates the accuracy of a nucleobase call. Specifically, a base call quality metric includes a value that indicates the likelihood that one or more predicted nucleobase calls for a genomic coordinate contain an error. For example, in certain implementations, a base call quality metric may include a Q score (e.g., PHil's Read EDitor (PHRED) quality score) that predicts the probability of an error for any given nucleobase call. To illustrate, a quality score (or Q score) may indicate that the probability of an incorrect nucleobase call at a genomic coordinate is equal to 1:100 for a Q20 score, 1:1000 for a Q30 score, 1:10000 for a Q40 score, and so on.

[0050] Relatedly, in some embodiments, the detection integration system can generate sequencing metrics by modifying or updating previous metrics in some embodiments. Such "reengineered sequencing metrics" may refer to sequencing metrics that have been updated, modified, amplified, refined or reengineered to measure or compare nuclear base calls (e.g., nuclear base calls for reads, genotypes or variant calls) relative to other nuclear base calls, standards or references, or for targeting specific goals or tasks. For example, the reengineered sequencing metrics may include modifications to the original sequencing metrics or a combination of original (e.g., unmodified) sequencing metrics. In some embodiments, for example, the detection integration system generates one or more of sequencing metrics based on reads, sequencing metrics from external sources, and / or sequencing metrics generated by the detection model as reengineered sequencing metrics. In some cases, the reengineered sequencing metrics refer to sequencing metrics generated by the detection integration system and are therefore proprietary or internal to the detection integration system and unavailable to third-party systems. Exemplary reengineered sequencing metrics include a comparative mapping quality distribution metric indicating a comparison between a mapping quality distribution associated with a reference sequence and alternative supporting nucleotide reads or a comparative base quality metric indicating a comparison between base qualities of a reference sequence and alternative supporting nucleotide reads.

[0051] As mentioned above, the detection integration system can use a machine learning model to modify the sequencing metrics and update the base calls. As used herein, the term "machine learning model" refers to a computer algorithm or a collection of computer algorithms that automatically improves a specific task through experience based on data usage. For example, a machine learning model can use one or more learning techniques to improve accuracy and / or effectiveness. Example machine learning models include various types of decision trees (e.g., gradient boosting trees), support vector machines, Bayesian networks, or neural networks.

[0052] In some cases, the detection integration system uses a genotype detection integration machine learning model to generate, modify or update predictions for genotype detection based on sequencing metrics. As used herein, the term "genotype detection integration machine learning model" refers to a machine learning model that generates predictions (such as genotype probabilities and / or variant detections) for one or more genomic samples. As indicated above, the genotype detection integration machine learning model includes a machine learning model that generates predictions for genotype detection of one or more genomic samples based on data from different types of nucleotide reads. For example, in some cases, the genotype detection integration machine learning model is trained to generate genotype probabilities indicating the probability or possibility of various genotypes at one or more genomic coordinates based on sequencing metrics. For example, the genotype detection integration machine learning model is trained to generate variant detection classifications indicating various probabilities or predictions for variant detection based on sequencing metrics. In some cases, the genotype detection integrated machine learning model is a series of gradient boosted decision trees (e.g., XGBoost algorithm or Treelite algorithm for decision tree ensembles), while in other cases, the genotype detection integrated machine learning model is a random forest model, a multilayer perceptron, a linear regression, a support vector machine, a deep table learning architecture, a deep learning transformer (e.g., a self-attention-based table transformer), or a logistic regression. In certain embodiments, the genotype detection integrated machine learning model includes multiple sub-models or operates in conjunction with (an instance of) another genotype detection integrated machine learning model. For example, a first genotype detection integrated machine learning model (e.g., an ensemble of gradient boosted trees) generates a first set of predictions for a first variant type (e.g., SNP) at genomic coordinates, and a second genotype detection integrated machine learning model generates a second set of predictions for a second variant type (e.g., indels) at genomic coordinates.

[0053] Relatedly, the term "variant call classification" refers to a predictive classification from a genotype call integrated machine learning model that indicates a probability, score, or other quantitative metric associated with some aspect of a genotype call based on one or more sequencing metrics (and how the genotype call impacts the variant call). Depending on the application of the genotype call integrated machine learning model (such as for predicting insertions and deletions), the variant call classification may include specialized predictions. For example, the variant call classification may include, but is not limited to: (i) for one or more genomic coordinates of a genomic sample, the probability of a true positive variant that a genotype call constitutes a true positive variant; (ii) the probability of a zygosity error that a genotype call includes a genotype zygosity error at one or more genomic coordinates; or (iii) a reference probability of a homozygous reference genotype at one or more genomic coordinates. Therefore, the term "reference probability" may refer to the probability of a homozygous reference genotype occurring at one or more genomic coordinates. As explained below, in some cases, the genotype call integrated machine learning model generates a variant call classification based on a first type of nucleotide read (e.g., SBS read) and a second type of nucleotide read (e.g., assembled nucleotide read).

[0054] As further used herein, the term "genotype probability" refers to the possibility, probability or score of a specific genotype at a genomic coordinate or genomic region. For example, the genotype probability includes the possibility of a homozygous reference genotype at one or more genomic coordinates, the possibility of a heterozygous variant genotype, or the possibility of a homozygous variant genotype. In some cases, the genotype probability may refer to a posterior genotype probability. Therefore, in some cases, the genotype probability determined by the genotype detection integrated machine learning model may be presented in a posterior genotype probability (GP) field of a VCF (such as a merged VCF) (or modified to be presented therein). Depending on the application of the genotype detection integrated machine learning model (such as for predicting SNPs), the genotype probability may include a special prediction.

[0055] As described above, the detection integrated system can generate a genotype probability and / or variant detection classification indicating or reflecting the possibility of identifying a variant at a genomic coordinate. As used herein, the term "variant" refers to one or more nucleobases that are not aligned, different or changed with a corresponding nucleobase (or multiple nucleobases) in a reference sequence or reference genome. For example, variants include SNPs, insertions and deletions, or structural variants, which indicate nucleobases in a sample nucleotide sequence that are different from nucleobases in the corresponding genomic coordinates of a reference sequence.

[0056] As mentioned, in some embodiments, the detection integration system modifies the data fields corresponding to the variant detection files. As used herein, the term "variant detection file" refers to a digital file indicating or representing one or more nuclear base detections (e.g., variant detections) compared to a reference genome and other information related to these detections. In some cases, the variant detection file may also include a genotype detection of a genomic sample, which indicates a reference detection or variant detection of an allele at a specific genomic coordinate or region. For example, a variant detection format (VCF) file refers to a text file format containing information about variants at specific genomic coordinates, the text file format including a meta-information row, a header row, and a data row, wherein each data row contains information about a single nuclear base detection (e.g., a single variant). As further described below, the call integration system can generate different versions of variant call files, including a pre-filter variant call file that includes variant nucleobase calls that pass or fail a quality filter of a base call quality metric, or a post-filter variant call file that includes variant nucleobase calls that pass the quality filter but excludes variant nucleobase calls that fail the quality filter.

[0057] Relatedly, a "merged variant call file" refers to a variant call file generated from one or more other variant call files. For example, a merged variant call file refers to a variant call file generated by selecting or merging data from a variant call file associated with one or more genotype calls determined from a first type of nucleotide read and a variant call file associated with one or more genotype calls determined from a second type of nucleotide read. In some cases, the merged variant call file includes some data selected from an initial variant call file and other data selected from different initial variant call files. Additionally, the merged variant call file may include data from the merged position, wherein some fields are generated to include new data not found in other (e.g., unmerged) variant call files. Therefore, in some embodiments, a merged variant call file is generated based on initial variant call files associated with different types of nucleotide calls.

[0058] In some embodiments, the detection integration system modifies the data fields corresponding to the metrics of the nuclear base calls associated with the variant call files, such as the fields of call quality, genotype and genotype quality. As used herein, when used relative to the data fields in the variant call files, the term "call quality" refers to a measure or indication of the possibility or probability of the presence of a variant at a given position. Therefore, the call quality field (or QUAL field) corresponding to the VCF file may include base call quality metrics such as PHRED scaled quality or Q score, which represents the probability that the genomic coordinates of the sample genome include the variant. Similarly, when used relative to a field, "genotype quality" refers to the possibility or probability that the specific predicted genotype of the nuclear base call is correct.

[0059] As noted, in some embodiments, the detection integration system utilizes a detection generation model to generate the nuclear base detection of genomic coordinates. As used herein, the term "detection generation model" refers to a probability model for generating sequencing data based on the nucleotide reads of a sample nucleotide sequence, and the sequencing data includes nuclear base detection, variant detection and / or genotype detection and associated metrics. Therefore, in some cases, the detection generation model can be a variant detection generation model. For example, in some cases, the detection generation model refers to a Bayesian probability model for generating variant detection based on the nucleotide reads of the sample nucleotide sequence. Such a model can process or analyze sequencing metrics corresponding to read stacking (e.g., multiple nucleotide reads corresponding to a single genomic coordinate), including mapping quality, base quality, and various assumptions, including exogenous reads, missing reads, joint detection, and the like. The detection generation model can also include multiple components, including but not limited to different software applications or components for mapping and alignment, sorting, repeat marking, calculating read stacking depth, and variant detection. In some cases, the call generation model refers to the ILLUMINA DRAGEN model for variant calling functions and mapping and alignment functions (e.g., DRAGEN variant caller or "DRAGEN VC").

[0060] The following paragraphs describe the detection integration system with respect to illustrative drawings that illustrate example embodiments and implementations. For example, Figure 1 A schematic diagram of a system environment (or "environment") 100 in which a detection integration system 106 operates according to one or more embodiments is illustrated. As illustrated, the environment 100 includes one or more server devices 102 connected to a client device 108, a local device 116, and a sequencing device 114 via a network 112. Figure 1 An embodiment of the checkout integrated system 106 is shown, but the present disclosure describes alternative embodiments and configurations below.

[0061] like Figure 1As shown, server device 102, client device 108, local device 116, and sequencing device 114 can communicate with each other via network 112. Network 112 includes any suitable network over which computing devices can communicate. Fig.12 The example network is discussed in more detail.

[0062] like Figure 1 Indicated, sequencing equipment 114 includes equipment for sequencing nucleic acid polymers. In some embodiments, sequencing equipment 114 analyzes nucleic acid fragments or oligonucleotides extracted from genomic samples to directly or indirectly generate nucleotide reads or other data on sequencing equipment 114 using (described herein) computer-implemented methods and systems. More specifically, sequencing equipment 114 receives and analyzes the nucleic acid sequence extracted from genomic samples in nucleotide sample slides (e.g., circulation cells). In one or more embodiments, sequencing equipment 114 utilizes SBS to sequence nucleic acid polymers into nucleotide reads. As a supplement or alternative to communicating across network 112, in some embodiments, sequencing equipment 114 bypasses network 112 and communicates directly with client device 108.

[0063] like Figure 1 As further indicated, the local device 116 is located at or near the same physical location as the sequencing device 114. In fact, in some embodiments, the local device 116 and the sequencing device 114 are integrated into the same computing device. The local device 116 can operate the call integration system 106 to generate, receive, analyze, store, and transmit digital data, such as by receiving sequencing metrics or determining genotype calls and / or variant calls based on analyzing such sequencing metrics. Figure 1 As shown, the sequencing device 114 can send (and the local device 116 can receive) sequencing metrics generated during a sequencing run of the sequencing device 114. By executing software in the form of a call integration system 106, the local device 116 can align nucleotide reads to a reference genome and / or utilize a genotype call integration machine learning model 107 to determine genotypes and / or genetic variants based on the sequencing metrics. The local device 116 can also communicate with a client device 108. Specifically, the local device 116 can transmit data to the client device 108, the data including a variant call file (VCF), sequencing metrics, or other information indicating a nucleobase call, a genotype call, a variant call, a sequencing metric, error data, or other metrics.

[0064] like Figure 1 As further indicated, the server device 102 can generate, receive, analyze, store, and transmit digital data, such as data used to determine genotype calls or sequence nucleic acid polymers. Figure 1As shown, the sequencing device 114 can send (and the server device 102 and / or the local device 116 can receive) call data and / or sequencing metrics. The server device 102 can also communicate with the client device 108 and / or the local device 116. Specifically, the server device 102 and / or the local device 116 can transmit data to the client device 108, the data including variant call files, or other information indicating nucleobase calls, sequencing metrics, variant calls, sequencing metrics, error data, or other metrics.

[0065] In some embodiments, the server device 102 comprises a distributed server collection, wherein the server device 102 comprises a number of server devices distributed across the network 112 and located in the same or different physical locations. In addition, the server device 102 may include a content server, an application server, a communication server, a web hosting server, or another type of server. In some cases, the server device 102 is located at the same physical location as the sequencing device 114 and / or the local device 116.

[0066] like Figure 1 As further shown, the server device 102 and / or the sequencing device 114 may include a sequencing system 104. Typically, the sequencing system 104 analyzes read data and / or call data (such as sequencing metrics received from the sequencing device 114) to determine the nucleobase sequence of the nucleic acid polymer. For example, the sequencing system 104 may receive raw data from the sequencing device 114 and may determine the nucleobase sequence of the nucleic acid fragment. In some embodiments, the sequencing system 104 determines the sequence of the nucleobase in the DNA and / or RNA fragment or oligonucleotide. In addition to processing and determining the sequence of the nucleic acid polymer, the sequencing system 104 also generates a variant call file indicating one or more genotype calls and / or variant calls for one or more genomic coordinates.

[0067] As just mentioned, and as Figure 1As illustrated in , the detection integration system 106 analyzes the detection data (such as sequencing metrics from the sequencing device 114) to determine the genotype detection of the sample nucleotide sequence for the genomic sample. The detection integration system 106 includes a detection generation model and a genotype detection integration machine learning model 107. In some embodiments, the detection integration system 106 determines the sequencing metrics of the sample nucleotide sequence. Based on the data derived or prepared from the sequencing metrics, the detection integration system 106 trains and applies the detection generation model to determine the nuclear base detection for the sample sequence corresponding to the genomic coordinates. The detection integration system 106 further uses the genotype detection integration machine learning model 107 to generate a prediction set (e.g., genotype probability for SNP or variant detection classification for insertion and deletion) to update or modify the genotype detection (and / or variant detection). Based on this data, for example, the detection integration system 106 can update the data field corresponding to the variant detection file to update the genotype detection and / or variant detection, thereby improving accuracy.

[0068] like Figure 1 As further illustrated and indicated in , the client device 108 may generate, store, receive, and send digital data. Specifically, the client device 108 may receive sequencing metrics from the sequencing device 114. In addition, the client device 108 may communicate with the server device 102 and / or the local device 116 to receive a variant call file including genotype calls and / or other metrics (such as call quality and / or genotype quality). The client device 108 may accordingly present or display information related to genotype calls to a user associated with the client device 108 within a graphical user interface. For example, the client device 108 may present a contribution metric interface that includes visualizations or depictions of various contribution metrics associated with or attributed to individual sequencing metrics for specific nucleobase calls.

[0069] Figure 1 The client device 108 illustrated in the example may include various types of client devices. For example, in some embodiments, the client device 108 includes a non-mobile device, such as a desktop computer or a server, or other types of client devices. In other embodiments, the client device 108 includes a mobile device, such as a laptop, a tablet computer, a mobile phone, or a smart phone. Fig.12 Additional details regarding client device 108 are discussed.

[0070] like Figure 1As further illustrated in FIG. 1 , the client device 108 includes a sequencing application 110. The sequencing application 110 may be a web application or a native application (e.g., a mobile application, a desktop application) stored and executed on the client device 108. The sequencing application 110 may include instructions that, when executed, cause the client device 108 to receive data from the call integration system 106 and present the data from the variant call file for display at the client device 108. In addition, the sequencing application 110 may direct the client device 108 to display a visualization of a contribution measure of a sequencing metric for a genotype call.

[0071] like Figure 1 As further illustrated in , the detection integration system 106 can be located on the client device 108 as part of the sequencing application 110, or on the sequencing device 114, or on the local device 116. Therefore, in some embodiments, the detection integration system 106 is implemented by being located (e.g., completely or partially located) on the client device 108. In other embodiments, the detection integration system 106 is implemented by one or more other components of the environment 100, such as the sequencing device 114 or the local device 116. Specifically, the detection integration system 106 can be implemented in a variety of different ways across the server device 102, the network 112, the client device 108, and the sequencing device 114. For example, the detection integration system 106 can be downloaded from the server device 102 to the client device 108, the local device 116, and / or the sequencing device 114, wherein all or part of the functionality of the detection integration system 106 is executed at each corresponding device within the environment 100.

[0072] although Figure 1 The components of the environment 100 are illustrated as communicating via the network 112, but in some implementations, the components of the environment 100 may also communicate directly with each other, bypassing the network 112. For example, and as previously mentioned, in some implementations, the client device 108 communicates directly with the sequencing device 114 and / or the local device 116. Additionally, in some embodiments, the client device 108 communicates directly with the detection integration system 106 (hosted on one or more of the illustrated components). In addition, the detection integration system 106 can access one or more databases housed on or accessed by the server device 102 or located elsewhere in the environment 100.

[0073] As indicated above, the call integration system 106 can determine output genotype calls based on one or more sequencing metrics of initial genotype calls from different types of nucleotide reads. Specifically, the call integration system 106 can generate predictions (e.g., genotype probabilities or variant call classifications) based on the sequencing metrics using a genotype call integration machine learning model, and can determine or update various metrics associated with the genotype calls (e.g., within a VCF file) based on the generated predictions. According to one or more embodiments, Figure 2 An example overview of the call integration system 106 determining output genotype calls based on genotype probabilities or variant call classifications from a genotype call integration machine learning model is illustrated. Figure 2 Additional details of the action.

[0074] like Figure 2 As illustrated in , the detection integration system 106 performs action 202 to receive a first genotype call and a second genotype call. Specifically, in some embodiments, the detection integration system 106 receives a first genotype call indicated by a first VCF file generated from a nucleotide read of the first read type. In addition, the detection integration system receives a second genotype call indicated by a second VCF file generated from a nucleotide read of the second read type. In some cases, the detection integration system 106 generates a first genotype call by analyzing SBS reads (e.g., nucleotide reads synthesized from sample library fragments shorter than a first threshold number of nucleobases). In these or other cases, the detection integration system 106 generates a second genotype call by analyzing different types of read data, such as: (i) assembled nucleotide reads, i.e., nucleotide reads assembled from shorter nucleotide reads to form a continuous sequence, (ii) CCS reads, and / or (iii) nanopore long reads. In some embodiments, the first received genotype call and the second received genotype call are initial genotype calls that the call integration system 106 uses as a basis for ultimately generating an output genotype call (e.g., by merging data associated with the first genotype call and the second genotype call).

[0075] like Figure 2As further illustrated in , the detection integration system 106 performs action 204 to identify sequencing metrics. Specifically, the detection integration system 106 identifies or determines sequencing metrics, such as sequencing metrics based on reads, sequencing metrics from external sources, and sequencing metrics generated by detection models. For example, the detection integration system 106 determines sequencing metrics indicating various attributes or data related to various genotype calls of nucleotide reads from sample nucleotide sequences. In some embodiments, the detection integration system 106 determines or extracts different sequencing metrics for generating genotype calls associated with different variant types (such as SNPs and insertions and deletions). In fact, according to different sequencing metrics, the detection integration system 106 can generate output genotype calls corresponding to the corresponding variant types, and the genotype call integration machine learning model is trained on the output genotype calls.

[0076] In detail, Figure 2 As illustrated in , the detection integration system 106 uses different instances of the genotype detection integration machine learning model to generate different predictions for different variant types based on the extracted sequencing metrics. For example, in order to generate an output genotype detection corresponding to a (biallelic) SNP, the detection integration system 106 performs action 206 to generate a genotype probability. As another example, in order to generate an output genotype detection corresponding to an insertion and deletion (or a multi-allelic SNP or a variant type other than a biallelic SNP), the detection integration system 106 performs action 208 to generate a variant detection classification. As indicated below, in some embodiments, the detection integration system 106 may use one or both of the following: a SNP-specific genotype detection integration machine learning model to generate a genotype probability; and an insertion and deletion-specific genotype detection integration machine learning model to generate a variant detection classification. In some cases, the detection integration system 106 may use a biallelic SNP genotype detection integration machine learning model to analyze or determine a biallelic SNP. In these or other cases, the call integration system 106 can use an indel-specific genotype call integration machine learning model to analyze or determine indels, multi-allelic SNPs, or other variant types that are not biallelic SNPs.

[0077] To generate genotype probabilities (e.g., via action 206), the detection integration system 106 uses a genotype detection integration machine learning model to analyze sequencing metrics (e.g., SNP-related sequencing metrics). Specifically, the detection integration system 106 uses a genotype detection integration machine learning model trained with SNP training data to generate genotype probabilities for one or more candidate SNPs. Based on the sequencing metrics, the genotype detection integration machine learning model generates a genotype probability set for a specific genomic coordinate, which indicates the possibility of a 0 / 0 genotype (e.g., a homozygous reference genotype), the possibility of a 0 / 1 genotype or a 1 / 0 genotype (e.g., a heterozygous genotype), and the possibility of a 1 / 1 genotype (e.g., a homozygous alternative genotype).

[0078] In order to generate a variant call classification (e.g., via action 208), the call integration system 106 generates (or updates or refines) a variant call classification based on a sequencing metric using a genotype call integration machine learning model. In detail, the call integration system 106 processes or analyzes one or more sequencing metrics using a genotype call integration machine learning model and generates a classification set (e.g., a predicted probability associated with a variant, zygosity, or reference call). For example, the call integration system 106 generates a variant call classification set using a genotype call integration machine learning model, and the variant call classification set includes: i) a first true positive variant probability for a first genotype call (e.g., from a first read type), ii) a second true positive variant probability for a second genotype call (e.g., from a second read type), iii) a first zygosity error probability for a first genotype call, iv) a second zygosity error probability for a second genotype call, and v) a reference probability.

[0079] like Figure 2As further illustrated in , the detection integration system 106 also performs action 210 to generate an output genotype call. Specifically, based on the genotype probability output by the genotype call integration machine learning model, the detection integration system 106 generates an output genotype call for one or more genomic coordinates of the SNP. Additionally or alternatively, based on the variant call classification output by the genotype call integration machine learning model, the detection integration system 106 generates an output genotype call for one or more genomic coordinates of the insertion and deletion. For SNP or insertion and deletion, the detection integration system 106 determines or updates the genotype call by generating a merged VCF file, which merges the data associated with the first read type (e.g., SBS reads) and the data associated with the second read type (e.g., assembled nucleotide reads), or otherwise generates from the data associated with the first read type and the data associated with the second read type. In some cases, the detection integration system 106 determines an output genotype indicating the presence or absence of a SNP or insertion and deletion at one or more genomic coordinates of a genomic sample. For example, the call integration system 106 selects an initial genotype call (e.g., a first genotype call or a second genotype call) as an output genotype call. Alternatively, the call integration system 106 generates an output genotype call that is different from the initial genotype call (e.g., a first genotype call and a second genotype call) but is based on data associated with the initial genotype call.

[0080] In some embodiments, the call integration system 106 uses a call generation model to generate a merged VCF file (such as generated by a genotype call integration machine learning model) based on genotype probabilities and / or variant call classifications. For example, the call integration system 106 applies multiple Bayesian probability models or algorithms to derive various probabilities for different nucleobases, quality metrics, mapping metrics, joint metrics, and other data occurring within the sample nucleotide sequence to be included in the variant call file. Based on the probability model, the call integration system 106 can further determine an output genotype call that indicates a predicted genotype (or variant) of the sample genome at the genomic coordinates corresponding to the reference genome.

[0081] As part of generating an output genotype call, in certain embodiments, the call integration system 106 generates, recalibrates, determines, modifies, confirms, or amplifies an initial genotype call using genotype probabilities and / or variant call classifications. In detail, the call integration system 106 uses genotype probabilities and / or variant call classifications (and / or other features) to determine or update certain metrics associated with the genotype call. For example, the call integration system 106 modifies data fields corresponding to variant call files for metrics such as call quality, genotype, and genotype quality (or other metrics as described below) to generate an output genotype call (e.g., as a new genotype call or as a modified or merged version of a first genotype call and / or a second genotype call).

[0082] although Figure 2 The specific order of actions 202-210 is illustrated, but in some embodiments, the detection integration system 106 performs these actions in different orders and / or in series at the same time. For example, the detection integration system 106 may perform action 206 to generate genotype probability and / or perform action 208 to generate variant detection classification (while performing action 210 to generate output genotype detection or during this process). For example, the detection integration system 106 simultaneously implements the genotype detection integrated machine learning model and the detection generation model to generate output genotype detection and the genotype probability / variant detection classification for modifying the output genotype detection. In some cases, the detection integration system 106 further modifies the data field of the merged variant detection file corresponding to the output genotype detection (for example, in the pre-filter or post-filter variant detection file). As proposed above, this simultaneous or parallel operation provides improved computational efficiency and increased speed to the detection integration system 106 by recalibrating them when the genotype detection is initially generated (rather than performing another operation before one operation).

[0083] In one or more specific implementations, the detection integration system 106 determines the output genotype call as part of the genomic coordinates for the SNP or insertion and deletion marker. For example, the detection integration system 106 determines the output genotype call to represent the SNP at the genomic coordinates (e.g., chr1:151863125) by identifying a G in the sample nucleotide sequence (where A is present in the reference genome). As another example, the detection integration system 106 determines that the genotype call indicates a deletion around one or more genomic coordinates (e.g., chr1:49263256) by identifying a single G in the sample nucleotide sequence (where GTAAC is present in the reference genome). As another example, the detection integration system 106 determines that the sequence of the genotype call represents an insertion at the genomic coordinates (e.g., chr1:7602080) by identifying the TTTCC sequence in the sample nucleotide sequence (where T is present in the reference genome). In fact, in some cases, the insertion includes a genotype call sequence that replaces a single reference base at the genomic coordinates of the reference sequence.

[0084] As described above, in certain embodiments, the call integration system 106 receives, identifies, or determines initial genotype calls from different types of nucleotide reads. Specifically, the call integration system 106 utilizes a multi-read type pipeline to combine sequencing metrics or other data from one type of nucleotide reads (e.g., short reads or SBS reads) with sequencing metrics or other data from another type of nucleotide reads (e.g., long reads or assembled nucleotide reads) to generate output genotype calls from the initial genotype calls. Figure 3 1 illustrates example types of nucleotide reads that the call integration system 106 may analyze or receive data about as part of generating an output genotype call at a genomic coordinate according to one or more embodiments. As indicated above, in some embodiments, the call integration system 106 is based on Figure 3 The first type of nucleotide reads and the second type of nucleotide reads depicted in the figure are used to identify or determine genotype calls and corresponding sequencing metrics.

[0085] like Figure 3As illustrated in , the detection integration system 106 analyzes the read data associated with the first type of nucleotide read 302. Specifically, the detection integration system 106 receives or determines a first genotype call from the first type of nucleotide read 302. For example, the detection integration system 106 determines or receives a genotype call indicating a genotype or variant at a specific genomic coordinate, as indicated by the read of the first type of nucleotide read 302. In some embodiments, the first type of nucleotide read 302 includes short reads (e.g., shorter than a threshold length or composed of less than a threshold number of nucleobases), such as SBS reads synthesized from sample library fragments shorter than a threshold number of nucleobases. In certain embodiments, the detection integration system 106 determines the first type of nucleotide read 302 from a hole in a circulation pool and / or via a fluorescent label. In some cases, the detection integration system 106 utilizes cluster generation and SBS chemistry to sequence millions or billions of clusters in a circulation pool. During the SBS chemistry, for each cluster, the call integration system 106 stores the nucleobase calls from the nucleotide reads for each sequencing cycle via real-time analysis (RTA) software. Although the above-mentioned specific genomic coordinates include a first genotype call based on the reads of the first type of nucleotide reads 302 and a second genotype call based on the reads of the second type of nucleotide reads 304, some genomic coordinates include only genotype calls based on the reads of the first type of nucleotide reads 302 or only genotype calls based on the reads of the second type of nucleotide reads 304, but not both.

[0086] like Figure 3 As further illustrated in , the detection integration system 106 analyzes the read data associated with the second type of nucleotide read 304. Specifically, the detection integration system 106 receives or determines a second genotype call from the second type of nucleotide read 304. For example, the detection integration system 106 determines or receives a genotype call indicating a genotype or variant at a specific genomic coordinate, as indicated by the read of the second type of nucleotide read 304. More specifically, the second type of nucleotide read 304 may include long reads (e.g., reads longer than a threshold length or consisting of at least a threshold number of nucleobases), such as assembled nucleotide reads, CCS reads, and / or nanopore long reads.

[0087] Regarding the assembled nucleotide reads, the detection integration system 106 can determine the assembled nucleotide reads by utilizing a mutagenesis process and a rendering process. In detail, the detection integration system 106 can assemble, create, synthesize or generate long reads from short reads. For example, the detection integration system 106 can apply mutations to a set of short reads (e.g., SBS reads or other short reads) to introduce unique genetic signatures so that the assembly can work on low-complexity regions with multiple repeats. In some cases, the detection integration system 106 applies random mutations and uses the output of the mutated short reads to restore information in regions of the sample genome that are difficult to sequence using ordinary SBS technology. For example, the detection integration system 106 combines mutated short reads to form assembled long reads, and the detection integration system 106 further performs a rendering process to undo at least a portion of the mutation after the short reads are combined or assembled into long reads.

[0088] As described above, in certain described embodiments, the call integration system 106 determines or extracts sequencing metrics for genotype calls at genomic coordinates. Specifically, the call integration system 106 determines sequencing metrics based on calls corresponding to nucleotide reads from sample nucleotide sequences, such as sequencing metrics based on reads, sequencing metrics from external sources, and sequencing metrics generated by call models. FIG. 4A to FIG. 4C The detection integration system 106 according to one or more embodiments is illustrated to determine the sequencing metric. Specifically, Figure 4A The call integration system 106 is illustrated to determine a read-based sequencing metric based on a first type of nucleotide read and a second type of nucleotide read; Figure 4B The call integration system 106 is illustrated to determine a sequencing metric generated by a calling model for a genotype call corresponding to a first type of nucleotide read or a second type of nucleotide read; and Figure 4C The call integration system 106 is illustrated to identify or determine an externally sourced sequencing metric for genomic coordinates corresponding to genotype calls of a first type of nucleotide read or a second type of nucleotide read.

[0089] like Figure 4A As illustrated in the example, the detection integration system 106 accesses, retrieves, obtains, determines, receives or generates nucleotide reads, which include a first type of nucleotide read 402a (e.g., a first type of nucleotide read 302) and a second type of nucleotide read 402b (e.g., a second type of nucleotide read 304). For example, the detection integration system 106 uses a sequencing device 114 to determine nucleotide reads from a region of a sample nucleotide sequence (e.g., a sample genome). For example, the detection integration system 106 generates multiple nucleotide reads using sequencing by synthesis (SBS) technology, Sanger sequencing technology, assembled nucleotide read technology, or other sequencing technologies discussed herein to determine genotype calls for oligonucleotide clusters.

[0090] like Figure 4A As further illustrated in FIG. 1 , in some embodiments, the call integration system 106 performs read processing and mapping 404a for a first type of nucleotide read 402a, and performs read processing and mapping 404b for a second type of nucleotide read 402b. For example, the call integration system 106 uses RTA software to store base call data in the form of separate base call data files (or BCLs). In some cases, the call integration system 106 further converts the BCL files into sequence data 408a and 408b (e.g., via BCL to FASTQ conversion), such as Figure 4B , where sequence data 408a corresponds to a first type of nucleotide read 402a, and sequence data 408b corresponds to a second type of nucleotide read 402b.

[0091] like Figure 4A As shown, the detection integration system 106 generates multi-read coverage (e.g., read accumulation), which includes multiple nucleotide reads or core base detections corresponding to a single genomic coordinate. Specifically, in certain embodiments, the detection integration system 106 compares the nucleotide reads with the reference genome or receives information related to the read comparison. Specifically, the detection integration system 106 determines which (which) core bases of a given read are compared with which genomic coordinates of the reference sequence (or receives information indicating the comparison). Different reads have different lengths and include different core bases. Therefore, in some cases, the detection integration system 106 analyzes each nucleotide of each read to determine the position of the read "fitting" with respect to the reference sequence (or receives information indicating the position), such as the position of the base in the read and the base in the reference. In some cases, the detection integration system 106 compares many reads at a single genomic coordinate, thus causing the reads to pile up.

[0092] In certain embodiments, the detection integration system 106 performs additional statistical tests to determine or detect differences between metrics associated with the reference nucleotide sequence and metrics associated with alternative support nucleotide reads. Through these statistical tests, the detection integration system 106 redesigns the original sequencing metrics to determine a read-based sequencing metric 406a for the first type of nucleotide read 402a and a read-based sequencing metric 406b for the second type of nucleotide read 402b. In some embodiments, the detection integration system 106 determines a shared set of sequencing metrics associated with both the first type of nucleotide read 402a and the second type of nucleotide read 402b.

[0093] In some cases, the detection integration system 106 determines or extracts raw sequencing metrics, which include one or more of the following: (i) an alignment metric for quantifying the alignment of a sample nucleotide sequence with the genomic coordinates of an example nucleotide sequence (e.g., a reference genome or a nucleotide sequence from an ancestral haplotype), (ii) a depth metric for quantifying the depth of the nucleobase calls of the sample nucleotide sequence at the genomic coordinates of the example nucleotide sequence, or (iii) a detection quality metric for quantifying the quality of the nucleobase calls of the sample nucleotide sequence at the genomic coordinates of the example nucleotide sequence. For example, the detection integration system 106 determines a mapping quality metric (e.g., a MAPQ metric), a soft shearing metric, or other alignment metrics that measure the alignment of a sample sequence with a reference genome. For another example, the detection integration system 106 determines a forward-reverse depth metric (or other such depth metric) or a detectability metric (or other such detection quality metric) for genotype calls or variant calls.

[0094] As just mentioned, in some embodiments, the detection integration system 106 reengineers the original sequencing metrics to generate read-based sequencing metrics 406a and 406b, which are more informative for comparing metrics associated with a reference nucleotide sequence and metrics associated with various supporting alternative nucleotide reads. For example, the detection integration system 106 determines various metrics of the sample sequence related to the reference sequence, and further determines various metrics of the sample sequence related to alternative support sequences. In addition, the detection integration system 106 performs a comparative analysis between metrics associated with the reference sequence and metrics associated with alternative support reads.

[0095] For example, the detection integration system 106 compares the mapping of the nucleobases of the sample nucleotide sequence (e.g., the sample genome) to the reference sequence with the mapping of the nucleobases to various alternative support reads. In some cases, the detection integration system 106 determines the mapping quality associated with the reference sequence to compare with the mapping quality associated with the alternative support reads. For example, the detection integration system 106 determines mapping quality statistics that reflect the difference in the distribution of reads supporting the reference sequence and reads supporting alternative alleles.

[0096] In these or other cases, the detection integration system 106 determines mismatch counts between the sample sequence and the reference sequence and between the reference sequence and the alternative support reads. The detection integration system 106 further compares the mismatch counts to determine a comparative mismatch count metric. In addition, the detection integration system 106 determines a soft shearing metric of the sample sequence associated with the reference sequence, and further determines a soft shearing metric associated with the alternative support reads. The detection integration system 106 also compares the soft shearing metric between the reference sequence and the alternative support reads to generate a comparative soft shearing metric. Further, the detection integration system 106 compares base call quality metrics associated with the reference sequence and the alternative support reads and / or compares the query positions of the sample sequence associated with the reference sequence with those query positions associated with the alternative support reads.

[0097] like Figure 4A As further illustrated in , the detection integration system 106 utilizes comparison and / or other statistical tests to generate read-based sequencing metrics 406a and 406b. In some cases, the detection integration system 106 generates read-based sequencing metrics 406a and 406b to include one or more of the same metrics listed above. For example, based on a first type of nucleotide read 402a and a second type of nucleotide read 402b, the detection integration system 106 generates read-based sequencing metrics 406a and 406b, and these read-based sequencing metrics include: i) an allele frequency metric indicating the frequency of occurrence of an allele detected by a first genotype, an allele detected by a second genotype, or a different allele detected by an alternative genotype different from the first genotype and the second genotype, ii) a coverage metric indicating a specific (e.g., maximum or cumulative total) coverage depth of a first type of nucleotide read 402a corresponding to a first genotype detection or a second type of nucleotide read 402b corresponding to a second genotype detection. depth metric, iii) a mapping quality metric (e.g., a MAPQ score) of a first type of nucleotide read 402a corresponding to a first genotype call or a second type of nucleotide read 402b corresponding to a second genotype call, iv) a nucleobase composition metric indicating the composition or composition of nucleobases at the genomic coordinates of one or more nucleotide reads from the first type of nucleotide read 402a or the second type of nucleotide read 402b, and v) an average coverage depth metric indicating the average (e.g., mean or median) coverage depth of the first type of nucleotide read 402a corresponding to the first genotype call or the second type of nucleotide read 402b corresponding to the second genotype call.

[0098] Additionally, the detection integration system 106 utilizes comparison and statistical tests to generate read-based sequencing metrics 406b based on the second type of nucleotide reads 402b, which may not be applicable to the first type of nucleotide reads 402a. For example, the detection integration system 106 generates read-based sequencing metrics 406b, which include: i) an assembly score indicating the accuracy or completeness of the assembled reads generated using mutagenesis and rendering, ii) a k-mer statistic indicating the length of the read and / or the length of the variant (e.g., insertion or deletion), and iii) a rendering metric indicating the accuracy or completeness of the mutation rendered from the assembled nucleotide reads (e.g., from the mutagenesis process). Additional details about read-based sequencing metrics 406a and 406b are provided below.

[0099] A. Read-based sequencing metrics

[0100] The following paragraphs describe various sequencing metrics based on reads in more detail. For example, the detection integration system 106 determines the base detection quality score of the base detection in the nucleotide read. Specifically, the detection integration system 106 determines the probability of the correctness of the core base detection of the nucleotide read (e.g., PHRED+33 encoding). In some cases, the detection integration system 106 determines one or more base detection quality scores in the form of DRAGEN QUAL score or Q score for one or more core base detections. In addition, the detection integration system 106 determines the score of the nucleotide reads supporting the alternative continuous sequence from the reference genome. For example, the detection integration system 106 determines the number of nucleotide reads supporting (e.g., matching or alignment) the alternative continuous sequence of the reference genome and the number of nucleotide reads supporting the primary assembly in the reference genome. The detection integration system 106 further compares the above-mentioned numbers and determines the score to reflect the comparison.

[0101] In some cases, the call integration system 106 uses specific features to determine the score of the reads supporting the alternative continuous sequence, including: i) the alignment score associated with the reference genome, ii) the alignment score associated with the assembly of the alternative continuous sequence, iii) the mapping quality of the nucleotide read, and iv) the amount of overlap with the genomic region. In addition, the call integration system 106 can classify the reads based on the alignment of the reads according to the following categories: i) perfect alignment with the assembly of the alternative continuous sequence (e.g., satisfying the first alignment score threshold), ii) perfect alignment with the reference genome, iii) strong alignment with the assembly of the alternative continuous sequence (e.g., satisfying the second alignment score threshold but not satisfying the first alignment score threshold), iv) strong alignment with the reference genome (e.g., also satisfying the second alignment score threshold but not satisfying the first alignment score threshold), and v) no strong alignment, whether it is the assembly of the alternative continuous sequence or the reference genome (e.g., failing to meet the second alignment threshold associated with both the assembly of the alternative continuous sequence and the reference genome). Based on these five categories, the call integration system 106 can further determine a score for comparing each of these categories to determine the score of nucleotide reads that support the alternative contiguous sequence (e.g., the score of reads that overlap with the target genomic region) and the score of nucleotide reads that support the reference genome.

[0102] In addition, the detection integration system 106 can determine the number of segmented nucleotide reads from the nucleotide reads corresponding to the genotype detection or variant detection as a sequencing metric based on the reads. More specifically, the detection integration system 106 determines the number of nucleotide read fragments that are not continuously aligned with the primary assembly of the reference genome (or the number of aligned bases is less than a threshold) but contain alignments with two or more reference sequences within the reference genome. For example, the detection integration system 106 uses the detection generation model to determine the segmented read counts that support genotype detection. For heterozygous deletion detection, the large segmented read counts of some false positive cases exceed the true positive cases, and the coverage depth is also higher than expected. Therefore, the detection integration system 106 can generate segmented nucleotide read metrics based on the nucleotide reads that support genotype detection.

[0103] In some embodiments, the detection integration system 106 compares the split read evidence of the alternative allele supporting the forward and nucleotide reads and reverse nucleotide reads, respectively. If most of the evidence comes from the forward or reverse reads, this deviation may indicate that there is a system problem, especially when the read count is relatively high (e.g., greater than 10 nucleotide reads). The detection integration system 106 uses the forward and reverse read counts with perfect alignment scores with continuous sequences as sequencing metrics for the genotype detection integration machine learning model.

[0104] As described above, the detection integration system 106 can determine the coverage depth of the nucleotide reads corresponding to the initial structural variant calls as a sequencing metric based on the reads. For example, the detection integration system 106 determines the count or number of nucleotide reads overlapping the target genomic region corresponding to the variant identified as present or absent by the initial genotype call or the initial variant call. Therefore, the coverage depth can be represented by the raw count of nucleotide reads that overlap the target genomic region by at least a threshold number of nucleotide bases.

[0105] In addition, as part of the sequencing metrics based on reads, the call integration system 106 can determine additional genotype calls (e.g., variant calls) within a threshold number of base pairs of an initial genotype call (e.g., variant call) from a genomic sample. For example, the call integration system 106 determines variant calls, such as insertions or deletions within a threshold proximity of the initial variant call (e.g., within 200 base pairs). Therefore, the call integration system 106 can use a code to indicate the presence or absence of such additional variant calls, such as a binary code of 0 for absence and 1 for presence.

[0106] In some embodiments, the detection integration system 106 further determines the alignment of the continuous sequence corresponding to the nucleotide read with the reference sequence of the reference genome modified to include the variant corresponding to the initial genotype call as a sequencing metric based on the read. Specifically, the detection integration system 106 modifies the reference genome by changing the nucleotide bases to reflect the variants (such as SNPs and insertions and deletions in the flanking regions). In theory, the modified reference genome can be perfectly aligned with the alternative continuous sequence, which provides some training benefits for the genotype call integration machine learning model to accurately identify variants.

[0107] In order to modify the reference genome to include variants, the detection integration system 106 can perform various steps. Specifically, the detection integration system 106 can remove a part of the sequence corresponding to the missing region (e.g., the missing region of the missing variant) from the reference genome. In some cases, the detection integration system 106 replaces the relevant part of the reference sequence in the FAST-All (FASTA) file with the continuous sequence representing the relevant variant. Then, the detection integration system 106 can regenerate the hash table using the modified FASTA file. In addition, the detection integration system 106 can run the mapping and comparison components of the detection generation model on the modified reference genome. The detection integration system 106 can further rerun the different detector components of the detection generation model on the new mapping and comparison output.

[0108] For candidate variants whose evidence based on reads is below a threshold (e.g., there are less than 5 or 10 nucleotide reads supporting the detection of candidate variants), one method for finding missing reads is to modify the local reference sequence by replacing the local reference sequence with a continuous sequence representing the candidate variant. For true positive cases, when the reads are remapped with the modified reference genome, some nucleotide reads in the nucleotide reads that are incorrectly mapped / aligned to the primary assembly of the reference genome will be more likely to be correctly mapped with the continuous sequence representing the candidate variant, thereby increasing the depth of reads on the newly modified reference genome. Based on the new mapping, if the detection integration system 106 reruns the detection generation model, the detection generation model does not detect variants for true homozygous deletions or insertions for true heterozygous deletions. Additionally, for the continuous sequence representing the candidate variant, the depth of read coverage should be increased relative to the original primary assembly, which should result in more accurate variant detection. The possibility of achieving more accurate mapping can be estimated by aligning the read length fragments of the continuous sequence representing the candidate variant with the reference genome.

[0109] In some embodiments, the detection integration system 106 analyzes the flanking regions of the variants in the sample sequence (as detected by the detection generation model), wherein the flanking regions include base calls within a threshold proximity of the variant (e.g., within 200 base pairs). For example, the detection integration system 106 uses the detection generation model (e.g., DRAGEN VC) to determine the initial variant based on the initial genotype call, modifies the reference genome to include (a portion of) the continuous sequence reflecting the variant, and identifies the flanking regions of the threshold size of 200 base pairs on either side of the variant. The detection integration system 106 further analyzes the flanking regions (e.g., left and right) of the combined sequence to determine the presence or absence of the variant. In fact, the detection integration system 106 can quantify the extent (e.g., quantity, magnitude, and / or size) of single nucleotide polymorphisms (SNPs) and / or insertions or deletions (indels) based on the modified reference genome (e.g., a combined sequence of the reference genome and the continuous sequence).

[0110] In some cases, the interpretation of continuous sequences is sensitive to the scoring parameters and penalties within the Smith-Waterman algorithm. Therefore, in these or other cases, the detection integrated system 106 uses the deletion counts output from the Concise Idiosyncratic Gap Alignment Report (CIGAR) string of multiple scoring parameter sets to measure the sensitivity to the Smith-Waterman scoring parameters / penalties. The detection integrated system 106 can further use the maximum continuous deletion length and the sum of all deletions corresponding to the genomic region spanned by the breakpoint as a sequencing metric (e.g., a sequencing metric based on reads).

[0111] In some cases, the detection integration system 106 determines a read-based sequencing metric in the form of a missing length of a nucleotide base based on one or more soft-cut nucleotide reads. For example, the detection integration system 106 re-aligns the soft-cut fragments from the nucleotide reads to determine the missing length (or the length of different types of variants). In some embodiments, the detection integration system 106 re-aligns only the soft-cut portion of the read to provide an estimate of the length of the missing or some other variant. For example, the detection integration system 106 performs re-alignment only when the size of the soft-cut portion meets (e.g., is greater than) a threshold number of soft-cut bases (e.g., 10 soft-cut bases or 20 soft-cut bases).

[0112] Additionally, in some embodiments, the call integration system 106 determines or calculates the realignment offset of the soft cut segments (e.g., those segments that meet the length requirement) by: i) for the soft cut reads on the left side of the called variant, aligning the soft cut portion to the left of the current position / coordinate indicating the end of the soft cut, ii) for the soft cut reads on the right side of the called variant, aligning the soft cut portion to the right of the current position / coordinate indicating the start of the soft cut, iii) determining the distance in number of nucleotide bases between the alignment position / coordinate and the position of the soft cut based on the original mapping, iv) determining the left mode and the right mode for all distances determined via steps i) to iii), and v) determining the left realignment offset and the right realignment offset by determining the difference between the left mode and the deletion length determined by the call generation model (e.g., DRAGEN SV caller) and the difference between the right mode and the deletion length determined by the call generation model (e.g., DRAGEN SV caller), such as the number of nucleotide bases determined based on the variant length (alternative sequence length).

[0113] In addition, the detection integration system 106 may determine a read-based sequencing metric in the form of the number of nucleotide reads that exhibit a mapping quality metric that fails to meet a threshold mapping quality metric. In detail, the detection integration system 106 corrects the situation in which the true positive shows that the nucleotide reads with a low MAPQ score (i.e., below the threshold MAPQ) are still correctly mapped (although the local alignment may be incorrect). In some cases, the detection integration system 106 uses MAPQ as a soft weight to indicate the possibility of aligning with an alternative continuous sequence or a reference genome. The detection integration system 106 may further determine the count or number of reads with a mapping quality metric (e.g., MAPQ score) that fails to meet (or is below) a threshold mapping quality metric (e.g., MAPQ=10 or MAPQ=60 or a relative MAPQ threshold). In some cases, the detection integration system 106 determines or generates variant calls based on the number of reads with a low mapping quality metric. In certain embodiments, such as in the case of MAPQ=60, the detection integration system 106 further incorporates the XQ score to determine the extended range of the possibility of the variant. The call integration system 106 can determine and incorporate the standard deviation of XQ across locally mapped reads to improve the predictions of the genotype call integration machine learning model.

[0114] As further shown above, in some embodiments, the call integration system 106 also determines an insert size representing the length of a nucleotide read fragment corresponding to an initial genotype call or variant call determined by the call generation model. Specifically, the call integration system 106 determines the size or length (e.g., number of base pairs) of an insert (or other variant) within a genomic region (e.g., SV region) of a genomic sample.

[0115] In some cases, the detection integration system 106 determines a read-based sequencing metric in the form of a palindrome metric. For example, the detection integration system 106 analyzes a portion of the reference sequence corresponding to the target genomic region in which the variant is detected (e.g., by detecting a generation model). Specifically, if the reference sequence in such a target genomic region is a palindrome (or within a threshold percentage of a palindrome or within a threshold number of base pairs from a palindrome), the likelihood of a folding effect increases. Based on this analysis, the detection integration system 106 identifies or detects fragments or portions of a genomic sample (e.g., a subsequence of a read), which are within a threshold distance from each other (e.g., within 200 base pairs) and are palindromes (which may exhibit missing due to folding effects during base detection). The detection integration system 106 can determine or measure the distance or proximity of a segment of a palindrome metric (e.g., the number of separated base pairs). In some cases, the call integration system 106 further combines the permutation entropy with the palindrome metric such that palindromic matches (eg, a pair of segments that appear palindromic to each other) with higher permutation entropy increase the likelihood of a deletion (or some other variant).

[0116] In addition, in some embodiments, the detection integration system 106 determines a read-based sequencing metric in the form of a variant possibility or probability based on the insert size, which variant possibility or probability represents the ratio of the initial variant detection of one or more genomic coordinates to the reference detection. Specifically, assuming that there is no variant, there is a certain implicit insert size or fragment size. On the other hand, assuming that there is a variant, there are different implicit insert sizes or fragment sizes. Therefore, based on the mean and standard deviation of the fragment size, the detection integration system 106 can determine which one is more likely between the presence or absence of the variant. For example, in some embodiments, the detection integration system 106 determines the ratio of the initial variant detection of one or more genomic coordinates to the reference detection according to the following formula:

[0117]

[0118] Where N A is the number of reads showing evidence supporting the alternative allele, l R,k is the original estimated insert size corresponding to read k assuming the absence of the variant, is the new estimated insert size based on alignment with the assembly of alternative contiguous sequences, μ I is the mean insert size of the variants in the genomic sample, and σ I is the standard deviation of the insert sizes of variants in a genomic sample assuming a Gaussian distribution. In some cases, Affected by the orientation and alignment of the segmented reads relative to the candidate deletion (or other type of variant).

[0119] Based on the read orientation and alignment relative to the candidate variant genomic region, the call integration system 106 can subtract the length of the proposed variant (e.g., deletion) from the original insert size estimate (e.g., based on the reference mapping and alignment). When considering all nucleotide reads that provide evidence for alternative allele support, the call integration system 106 can determine a likelihood ratio based on the expected insert size across the read set (e.g., alternative to reference).

[0120] In some cases, The estimate is affected by the orientation of the segmented reads as evidence of a variant (e.g., deletion). Therefore, the detection integration system 106 adjusts the insert size estimate based on the read orientation (e.g., for forward and reverse situations). However, the continuous sequence does not usually match the reference flanking region. Therefore, the insert size calculation will depend on the read orientation and the starting position of the segmented read relative to the breakpoint after alignment with the continuous sequence. Additionally, the reference start provided in the BAM file (e.g., the genomic coordinates of the variant start) does not usually include the soft shearing portion of the nucleotide read, and since the insert size calculation uses the actual start of the read, the detection integration system 106 adjusts the reference start to take into account the amount of soft shearing bases.

[0121] In one or more embodiments, the detection integration system 106 determines the sequencing metric based on the read segment in the form of a confidence interval around the end breakpoint. Specifically, the detection integration system 106 uses the detection generation model to determine the confidence interval as a measure of the certainty of the breakpoint position. For example, the detection integration system 106 determines the range of reference coordinates corresponding to the breakpoints detected by the variant may be located. In some cases, the detection integration system 106 determines the range of reference coordinates to reflect the threshold percentile (e.g., the 95th percentile) in terms of the confidence interval.

[0122] In certain embodiments, the detection integration system 106 further determines additional or alternative sequencing metrics based on reads. For example, the detection integration system 106 determines the homology length as a sequencing metric based on reads. Specifically, the detection integration system 106 determines the length of the repeated nucleotide base sequence in the target genomic region of the variant and / or the length of the nucleotide base sequence with at least a threshold homology measure with other nucleotide base sequences (of similar length) in the target genomic region of the structural variant (e.g., HOMLEN=8GCTTGAAC GCTTAAAC GCTAGAAC GCTTGAAC GCTTGTAC, etc.). In some cases, the detection integration system 106 determines the length of the inserted nucleotide base sequence as a sequencing metric based on reads. In these or other cases, the detection integration system 106 determines the homology of the inserted nucleotide base sequence relative to the reference sequence in the target genomic region of the variant.

[0123] In one or more embodiments, the detection integration system 106 determines additional or alternative read-based sequencing metrics, which include: i) a comparative mapping quality distribution metric that indicates a mapping quality distribution that compares the mapping quality associated with the reference sequence and the mapping quality associated with the alternative support read; ii) a comparative secondary mapping alignment metric that indicates a comparison between secondary mappings associated with bases in the reference sequence and bases in the alternative support read; iii) a comparative mismatch count metric that indicates a comparison between mismatched nucleobases associated with the reference sequence and mismatched bases associated with the alternative support read; iv) a comparative soft clipping metric that indicates a comparison between a soft clipping metric associated with the reference sequence and a soft clipping metric associated with the alternative support read; v) one or more comparative read depth metrics that indicate a read depth of a nucleotide read and one or more average reads. The invention relates to a method for detecting a base sequence of a nucleic acid sequence in a plurality of nucleotide reads, wherein the base quality of the nucleic acid sequence is compared with the base quality of the nucleic acid sequence in the plurality of nucleotide reads, and the base quality of the nucleic acid sequence is compared with the base quality of the nucleic acid sequence in the plurality of nucleotide reads. The method comprises the following steps:

[0124] B. Sequencing metrics generated by calling models

[0125] In addition to the read-based sequencing metrics 406a and 406b, Figure 4B As illustrated in FIG. 1 , the detection integration system 106 also generates the detection model generated sequencing metrics 412a and 412b. Specifically, the detection integration system 106 uses the detection model instances 410a and 410b to generate the detection model generated sequencing metrics 412a and 412b based on the sequence data 408a and 408b, respectively. For example, the detection integration system 106 generates the detection model generated sequencing metrics 412a and 412b based on the sequence data 408a and 408b, respectively. Figure 4AThe described read processing and mapping 404a extracts or determines sequence data 408a. Similarly, the detection integration system 106 extracts or determines sequence data 408b based on the read processing and mapping 404b. In some cases, the detection integration system 106 generates sequence data 408a and 408b as part of one or more digital files (such as BCL files and FASTQ files).

[0126] To generate such files, in some embodiments, the sequencing device 114 (or the detection integrated system 106) uses cluster generation and SBS chemical reactions to sequence millions or billions of clusters in the circulation pool. During the SBS chemical reaction, for each cluster, the sequencing device 114 (or the detection integrated system 106) stores the nuclear base calls of the first type of nucleotide reads 402a and the second type of nucleotide reads 402b for each sequencing cycle via real-time analysis (RTA) software. The sequencing device 114 (or the detection integrated system 106) further stores the base call data in the form of a separate base call data file (or BCL) using RTA software. In some cases, the sequencing device 114 (or the detection integrated system 106) (e.g., via BCL to FASTQ conversion) further converts the BCL file into sequence data 408a and 408b. For example, the sequencing device 114 (or the integrated sequencing system 106) generates a FASTQ file based on the first type of nucleotide reads 402a and the second type of nucleotide reads 402b, wherein the FASTQ file includes sequence data 408a and 408b, respectively.

[0127] In some cases, the call integration system 106 generates sequence data 408a and 408b for each cluster that passes the initial quality filter from the sample sequence. For example, the call integration system 106 generates an entry for each cluster, where each entry includes four rows (or four items of sequence data): i) a sequence identifier with information about the sequencing run and cluster; ii) the nucleobase calls that make up the sequence (e.g., a sequence of A calls, C calls, T calls, G calls, and / or N calls); iii) a separator (e.g., a "+" sign); and iv) a base call quality metric that indicates the probability (PHRED+33 encoding) of the correctness of the nucleobase call.

[0128] like Figure 4BAs further illustrated in , the detection integration system 106 implements, utilizes or applies the detection generation model 410a to process or analyze the sequence data 408a. Similarly, the detection integration system 106 implements, utilizes or applies the detection generation model 410b to process or analyze the sequence data 408b. In fact, in some embodiments, the detection integration system 106 generates the detection model-generated sequencing metrics 412a and 412b by reengineering the original sequencing metrics (e.g., the original sequencing metrics within the sequence data 408a and 408b) using the corresponding instances 410a and 410b of the detection generation model. Specifically, the instances 410a and 410b of the detection generation model include mapping and alignment components to map and align the nucleobase calls from the sequence data 408a and 408b. In addition, instances 410a and 410b of the call generation model include variant calling components to generate initial genotype calls (e.g., reference base calls, such as nucleobase calls, variant calls, or non-variant calls) based on the sequence data 408a and 408b. In some cases, the call integration system 106 extracts sequencing metrics 412a and 412b generated by the call generation model that have been generated using the mapping and alignment components and variant calling components of the instances 410a and 410b of the call generation model.

[0129] To illustrate examples of sequencing metrics 412a and 412b generated by the call model, in some cases, the call integration system 106 generates variant call metrics that include one or more of the following: i) genotype metrics that correspond to the GT field of the VCF file and indicate the genotype of the genomic coordinate, ii) base call quality metrics (e.g., DRAGEN QUAL scores) that indicate the quality score of the genotype call generated via the call generation model 410a or 410b, iii) genotype quality metrics (e.g., GQ scores) that indicate a measure of the confidence or quality of the predicted genotype for the genomic coordinate, iv) genotype probability metrics that indicate one or more probabilities of various genotypes occurring at the genomic coordinate, v) PHRED scaled likelihood metrics or non-PHRED scaled likelihood metrics that indicate the probability of error associated with the genotype call, vi) base call quality metrics (e.g., DRAGEN QUAL scores) that indicate the quality score of the genotype call generated via the call generation model 410a or 410b, a foreign read detection metric (e.g., a foreign read detection (FRD) score) generated by the base model, which indicates the probability that one or more of the first type of nucleotide read 402a or the second type of nucleotide read 402b in the pile is likely to be a foreign read (e.g., their true position is elsewhere in the reference sequence), vii) a base quality degradation metric (e.g., a base quality degradation (BQD) score) generated by the call model, which is based on the strand bias, the error position in the line, or the first type of nucleotide read 402a and / or One or more of the low mean base qualities of the subset of the second type of nucleotide reads 402b indicates a probability of decreased base quality, viii) an average read depth, ix) a normalized read depth, x) a read depth with mapq0 reads, xi) a read depth without mapq0 reads, xii) an indel statistic (e.g., a polymerase chain reaction or "PCR" curve), and / or xiii) a hidden Markov model (HMM) statistic, xiv) a secondary alignment metric indicating a probability that the secondary genotype call is correct, xv) a base background metric indicating a base background metric. Metrics indicating context information of nucleotides surrounding a genotype call, xvi) nearby call metrics indicating proximity to a genotype call (e.g., adjacent to a genotype call or within a threshold separation from a genotype call), xvii) joint detection metrics indicating the probability of a detection corresponding to the union of two or more overlapping genotype calls, and / or xviii) read filtering metrics indicating a threshold quality metric or other metric for filtering out genotype calls with low mapping quality, base quality, or other quality metrics; or other variant call metrics.The call integration system 106 generates call model-generated sequencing metrics 412a and 412b based on internal (e.g., proprietary and model-specific) variables that reflect interacting processing paths, corner cases, and difficult predictions / decisions.

[0130] As indicated above, in some cases, the detection integration system 106 determines the FRD score according to the method described in U.S. Patent Application No. 16 / 280,022, entitled "System and Method for Correlated Error Event Mitigation for Variant Calling" to Eric Jon Ojard, filed on February 19, 2019, which is incorporated herein by reference in its entirety. In certain specific implementations, the detection integration system 106 also (or alternatively) determines the BQD score, FRD score, HMM statistics, and / or other variant detection metrics according to the methods described in U.S. Patent Application Nos. 17 / 165,828, 15 / 643,381, and 14 / 811,836, which are incorporated herein by reference in their entirety.

[0131] like Figure 4BAs illustrated in FIG. 4 , the sequencing metrics 412 a and 412 b generated by the call generation model include, but are not limited to, variant call metrics extracted via the variant call components of the instances 410 a and 410 b of the call generation model. In addition to or in lieu of the examples of sequencing metrics 412a and 412b generated by the calling models described above, in some cases, the calling integration system 106 (e.g., via metric reengineering) generates variant calling metrics that include one or more of the following: i) the number of samples in a population; ii) the number of reads processed to generate genotype calls, the number of variants (e.g., SNPs and indels); iii) the number of bi-allelic sites (e.g., genomic coordinates containing two observed alleles); iv) the number of multi-allelic sites (e.g., the number of sites in a variant calling file containing three or more observed alleles); v) the number of SNPs; vi) the number of different types of indels (e.g., homozygous insertions, heterozygous insertions, and heterozygous deletions); vii) the total number of heterozygous indels (e.g., indels, indels, or deletions); viii) the total number of de novo (deletions); and viii) the number of genotypes. novo) SNPs (e.g., SNPs with a de novo quality metric that meets a threshold level); ix) the number of de novo indels (e.g., indels with a de novo quality metric that meets a threshold level); x) the number of de novo MNPs (e.g., MNPs with a de novo quality metric that meets a threshold level); xi) the number of SNPs in a first chromosome divided by the number of SNPs in a second chromosome; xiiv) the number of SNP transitions; xiii) the number of SNP transversions; xiv) the number of heterozygous variants; xv) the number of homozygous variants; xvi) the ratio between the number of heterozygous variants and the number of homozygous variants; xvii) the number of variants detected within the dbSNP reference file; and / or xviii) the total number of variants minus the number detected within the dbSNP file.

[0132] Additionally, the sequencing metrics 412a and 412b generated by the call model may include mapping and alignment sequencing metrics extracted via the mapping and alignment components of the call generation model 410a or 410b. For example, the call integration system 106 (e.g., via metric reengineering) generates or extracts mapping and alignment metrics that include one or more of the following: i) the number of total input reads; ii) the number of duplicate marker reads; iii) the number of duplicate marker reads and paired reads removed; iv) the number of unique reads; v) the number of reads with sequencing pairs; vi) the number of reads without sequencing pairs; vii) an indication of reads that failed quality checks; viii) an indication of mapped reads; ix) the number of unique and mapped reads; x) the number of unmapped reads; xi) the number of single reads (e.g., where a read is mapped but a paired read cannot be read); xii) the number of paired reads; xiii) the number of properly paired reads (e.g., where a pair Both reads are mapped and fall within an acceptable range of each other based on an estimated insert length distribution); xiv) the number of discordant reads (e.g., reads that are not properly paired); xv) the number of paired reads that map to different chromosomes; xvi) the number of paired reads that map to different chromosomes that also have a mapping quality metric of 10 or greater; xvii) the percentage of reads within R1 and R2 with indels; xviii) the percentage of soft-clipped bases in R1 and R2; xix) the number of mismatched bases in R1 and R2; xx) the number of bases (e.g., total and / or in R1 or R2) with a base quality of at least 30; xxi) the number of alignments (e.g., total alignments, secondary alignments, and / or supplemental alignments); xxii) estimated read length; and xxiiiv) estimated sample contamination.

[0133] C. Sequencing metrics from external sources

[0134] Now turn to Figure 4C , the detection integration system 106 generates, extracts or determines the sequencing metrics 416 from the external source. Specifically, the detection integration system 106 determines the sequencing metrics 416 from the external source based on one or more databases external to the detection integration system 106, such as the sequencing information database 414. For example, the detection integration system 106 accesses sequencing metrics that are universal or generally applicable to sequenced nucleotides. In addition, the detection integration system 106 accesses or determines sequencing information about a specific reference sequence (e.g., stored in the sequencing information database 414).

[0135] In some cases, the call integration system 106 determines sequencing metrics 416 from external sources, which include: i) mappability metrics that indicate the ease with which a particular nucleotide sequence (or a particular nucleotide read or nucleobase call) can be mapped to one or more genomic coordinates within a reference genome, ii) guanine-cytosine-content metrics that indicate the count (or loss or mean) of guanine-cytosine content in a reference nucleotide sequence (e.g., a reference genome), iii) replication timing metrics that indicate the time required to replicate a particular number of nucleotides from a reference sequence, iv) one or more DNA structure metrics that indicate the DNA structure of a reference sequence (e.g., a reference genome), A structure, v) a conservation metric indicating a measure of sequence conservation across multiple species (e.g., a measure of variation relative to a mean), vi) a confidence classification indicating the degree to which the nucleobase at one or more genomic coordinates can be accurately determined, vii) a repeat classification indicating the category of the repeated genomic region of the one or more genomic coordinates, viiii) a cytosine quadruplex indicator indicating that the one or more genomic coordinates are part of a cytosine quadruplex, ix) a guanine quadruplex indicator indicating that the one or more genomic coordinates are part of a guanine quadruplex, and / or x) a homopolymer indicator indicating that the one or more genomic coordinates are part of a homopolymer within a reference genome.

[0136] In some embodiments, the detection integration system 106 determines the sequencing metrics 416 of the external source by analyzing one or more genomic regions of the reference genome corresponding to (or aligned) one or more genomic coordinates detected with the initial genotype. Many challenging variant detections occur in the low-complexity genomic regions of the reference genome. In some cases, these genomic regions are characterized by multiple instances of long repeats (e.g., more than 50 base pairs), a certain combination of very large amounts (e.g., more than 10) of shorter repeats (e.g., 4 to 8 repeated bases), and sometimes contain a subset of bases (e.g., A and T, but not C or G). The nucleotide reads correctly aligned with such low-complexity genomic regions generally have a portion or fragment of the nucleotide reads mapped to the more unique sequences flanking the heavy repeat region. Alternatively, the reference genome or genomic sample may include some intermediate breaks (e.g., single bases between primary repeat patterns that destroy repetitiveness), which contribute to the alignment of nucleotide reads with the low-complexity genomic regions of the reference genome. However, when combined with SNPs, indels, and sequencing errors, alignment and read collection with sufficient evidence to compare reference to alternative allele support becomes problematic. Therefore, in some embodiments, the call integration system 106 monitors external sources of sequencing metrics 416 (associated with complexity) that can be augmented with read-based sequencing metrics to provide an overall assessment of the likelihood of a variant being present (for both Bayesian and machine learning approaches).

[0137] For example, the detection integration system 106 accesses or determines sequencing information about a specific reference genome (e.g., stored in a sequencing information database 414). In some cases, the detection integration system 106 determines sequencing metrics 416 from external sources, which include tandem repeat lengths in nucleotide bases of a target genomic region within a reference genome corresponding to a candidate region of a genomic sample. Specifically, the detection integration system 106 analyzes portions of a reference genome corresponding to a variant region of a genomic sample to identify tandem repeats (e.g., a sequence of two or more bases repeated multiple times in a head-to-tail manner) and further determines the length within the tandem repeats (e.g., the number of base pairs).

[0138] In certain embodiments, the detection integration system 106 determines the sequencing metric of the external source in the form of a repeatability metric or a homopolymer metric. In fact, an indicator of the possibility of an error mapping (e.g., an error mapping that results in a false positive) that needs to be corrected is based on the repeatability of bases within the reference sequence. Therefore, the detection integration system 106 can measure this repeatability using various sequencing metrics, which include: i) the maximum repeat pattern length, which indicates the maximum length of a base sequence repeated at least twice over the span of the (reference genome) of the candidate region (corresponding to), ii) the maximum repeat length percentage, which indicates the percentage of the region (corresponding to the portion of the reference genome of the region) consumed or occupied by the maximum repeat pattern length, and iii) the maximum homopolymer length, which indicates the length of the longest sequence of the same base in the (reference genome portion) of the candidate region (corresponding to).

[0139] In addition to or in lieu of the repeatability metric, in some cases, the detection integrated system 106 determines an externally derived sequencing metric in the form of permutation entropy of nucleotide bases. For example, the detection integrated system 106 determines a measure of randomness of a nucleotide sequence that can predict mapping / alignment accuracy. In some cases, the detection integrated system 106 determines permutation entropy by determining the entropy on the permutation of a nucleotide sequence of a given length. For example, the detection integrated system 106 can determine the permutation entropy according to the following formula:

[0140] S1∈{A,C,G,T}

[0141] S2∈{AA,AC,AG,AT,CA,CC,CG,CT,GA,GC,GG,GT,TA,TC,TG,TT}

[0142] S3∈{AAA,AAC,AAG,AAT,ACT,...,TTA,TTC,TTC,TTT}

[0143] S4∈{AAAA,AAAC,AAAG,AAAT,AACA,...,TTGT,TTTA,TTTC,TTTG,TTTT}

[0144] Where S N is the set of all permutations of a sequence of length N, where:

[0145] |S N |=4 N

[0146] So that the set S N The permutation element s appears in N,k The probability of is given by:

[0147]

[0148] where c k is the permutation element s in a sequence of length M N,k In some cases, the detection integration system 106 normalizes the permutation entropy to:

[0149]

[0150] in is a set of indices such that p N,k >0.

[0151] As described above, the detection integration system 106 can further determine the sequencing metrics of the external source in the form of identifying the presence or absence of cytosine quadruplexes (C-quadruplexes) or guanine quadruplexes (G-quadruplexes) in the target genomic region. In detail, the detection integration system 106 determines the counts of cytosine detection and guanine detection in the target genomic region of the reference genome, and the target genomic region of the reference genome corresponds to the variant region of the genomic sample or the genomic region considering the initial variant detection. In order to identify the cytosine quadruplex, the detection integration system 106 identifies the occurrence of four or more instantiations of three consecutive cytosine bases separated by one or more different nucleotide bases (in the target genomic region) (e.g., a pattern of CCC A CCC A CCC A CCC). Similarly, in order to identify the guanine quadruplex, the detection integration system 106 identifies the occurrence of four or more instantiations of three consecutive guanine bases separated by one or more different nucleotide bases (in the target genomic region) (e.g., a pattern of GGG T GGG T GGG T GGG).

[0152] In one or more embodiments, the detection integration system 106 identifies C-quadruplexes or G-quadruplexes, wherein up to a threshold number of nucleotide bases (e.g., up to 7 nucleotide bases) appear between instantiations of three Cs or three Gs. For example, the detection integration system 106 identifies GGG TACC GGG TGTACA GGG AAGTCT GGG as a G-quadruplex. In some cases, G-quadruplexes (and C-quadruplexes) are known to cause sequencing problems. Therefore, the detection integration system 106 uses the presence of such sequences to adjust the confidence in the mapping and alignment of the reads and the accuracy of the subsequent continuous sequence construction.

[0153] In certain embodiments, the detection integration system 106 determines a data compression metric as part of the sequencing metric 416 of the external source. Specifically, the detection integration system 106 uses one or more data compression algorithms to determine a data compression metric that quantifies the randomness of the sequence. One such compression algorithm for lossless compression is the Liv-Zempel-Welch algorithm. Using this algorithm, the detection integration system 106 constructs a dictionary of unique K-mers starting with length 1, and provides encoding for each entry in the dictionary. The detection integration system 106 can use the number of keys in the dictionary for structural variants and flanking regions in the reference genome as a sequencing metric.

[0154] In addition to or in lieu of the externally sourced sequencing metrics 416 described above, in some embodiments, the call integration system 106 determines a structural variant sequence alignment metric as part of the externally sourced sequencing metrics 416. For example, the call integration system 106 uses an ungapped alignment score and a Smith-Waterman alignment score for the proposed missing sequence for the left / right flanking genomic region in the reference. If there are multiple alignments with scores above a threshold ungapped alignment score and / or a threshold Smith-Waterman alignment score, the genotype call integration machine learning model may process the variant sequence alignment metric as an indicator of a higher likelihood of an inaccurate variant call.

[0155] In addition, the detection integration system 106 can also determine simulated read alignment metrics as sequencing metrics from external sources. Assuming that the continuous sequence representing or including the variant is accurate, there should theoretically be many nucleotide reads that have good alignments with the continuous sequence, even for heterozygous deletions. However, for low-evidence true positive cases of variants, there is a possibility of missing reads because the reads corresponding to the SV region are either mapped elsewhere or not mapped. Therefore, the detection integration system 106 can determine the possibility of missing reads by simulating reads.

[0156] Specifically, the detection integration system 106 selects a segment from a continuous sequence whose length is equal to the SBS read. The detection integration system 106 selects a segment of a continuous sequence that spans the breakpoint, is equal to the SBS read length, and is aligned to the reference sequence in the SV region. For cases where the alignment is ambiguous, the alternative alignment score will be higher and can be used as a possible guide for the expected read depth. The detection integration system 106 can further use a fragment of a continuous sequence that is equal to the read length symmetrical about the breakpoint to obtain the highest alignment score. The detection integration system 106 can further determine an additional offset from this symmetry point to check the alternative alignment score of the overlapping range.

[0157] In one or more embodiments, the detection integration system 106 determines additional or alternative sequencing metrics, including sequencing metrics based on reads, sequencing metrics generated by detection models, and / or sequencing metrics from external sources. For example, the detection integration system 106 determines the sequencing metrics in the following table, wherein each of these metrics belongs to one or more of sequencing metrics based on reads, sequencing metrics generated by detection models, and / or sequencing metrics from external sources.

[0158]

[0159]

[0160]

[0161] As described above, in certain described embodiments, the detection integration system 106 uses the above-mentioned sequencing metrics to generate a collection of machine learning predictions for different variant types. Specifically, the detection integration system 106 uses the genotype detection integration machine learning model to generate genotype probabilities (for SNPs) or variant detection classifications (for insertions and deletions) corresponding to various genomic coordinates. In addition, the detection integration system 106 determines the output genotype detection by generating a variant detection file (e.g., a merged variant detection file) based on the genotype probability and / or variant detection classification. According to one or more embodiments, FIG. 5A to FIG. 5C The call integration system 106 is illustrated as generating one or both of genotype probabilities and variant call classifications, generating genotype calls based on such probabilities and / or classifications, and generating a merged variant call file including genotype calls based on such probabilities and / or classifications. For example, Figure 5A The call integration system 106 according to one or more embodiments is illustrated using a genotype call integration machine learning model to generate genotype probabilities for (biallelic) SNPs based on sequencing metrics corresponding to initial genotype calls from different read types. Figure 5B The call integration system 106 according to one or more embodiments uses a genotype call integration machine learning model to generate a variant call classification for an insertion or deletion (or a multi-allelic SNP or a variant type other than a biallelic SNP) based on sequencing metrics corresponding to initial genotype calls from different read types. Thereafter, Figure 5C The call integration system 106 is illustrated as generating a variant call file including output genotype calls based on genotype probabilities and / or variant call classifications according to one or more embodiments.

[0162] like Figure 5AAs illustrated in , the call integration system 106 identifies the genomic coordinates 502. For example, the call integration system 106 identifies the genomic coordinates 502 based on the nucleobase calls corresponding to the sample nucleotide sequence or based on the haplotype data corresponding to the genomic coordinates 502. In some cases, the call integration system 106 identifies the genomic coordinates 502 by determining (i) one or more nucleobase calls from nucleotide reads covering the genomic coordinates and (ii) one or more nucleobase calls satisfy one or more threshold sequencing metrics (e.g., a base call quality metric of Q30). Additionally or alternatively, in certain embodiments, the call integration system 106 identifies the genomic coordinates 502 by based on a database including a haplotype reference panel associated with a particular genomic coordinate. Regardless of the identification method, in some cases, the call integration system 106 uses a call generation model 503 (e.g., a variant detector as part of the call generation model) to identify the genomic coordinates 502.

[0163] like Figure 5A , the call integration system 106 also generates initial genotype calls 505 using a call generation model 503. In detail, the call integration system 106 generates initial genotype calls 505 using a call generation model 503 (e.g., a DRAGEN detector) to predict the presence (or absence) of a variant (or a specific genotype) at a genomic coordinate 502. As described, the call generation model 503 generates initial genotype calls 505 by analyzing or processing sequencing metrics 504 (or a subset of sequencing metrics 504, such as sequencing metrics based on reads and sequencing metrics from external sources). In addition, the call generation model 503 also generates some sequencing metrics in the sequencing metrics 504 (e.g., sequencing metrics generated by the call model) as part of predicting the initial genotype calls 505.

[0164] In practice, the call integration system 106 determines a sequencing metric 504 for the genomic coordinate 502. Specifically, the call integration system 106 determines a sequencing metric associated with the nucleotide read, generated by the call generation model 503, or retrieved from an external source, as described above. Based on the sequencing metric 504, the call integration system 106 further generates a genotype probability 508, which together can indicate a measure of confidence or probability that the genomic coordinate 502 includes or exhibits a SNP variant.

[0165] Specifically, Figure 5AAs shown, the call integration system 106 utilizes the genotype call integration machine learning model 506 to generate genotype probabilities 508. For example, the genotype call integration machine learning model 506 analyzes or processes the sequencing metrics 504 and the initial genotype calls 505 as input to generate genotype probabilities 508 as outputs, which genotype probabilities include: i) a first genotype probability 510 that the initial genotype call 505 is a homozygous reference genotype at the genomic coordinate 502 (e.g., "L(0 / 0)@chr5:4"), ii) a second genotype probability 512 that the initial genotype call 505 is a heterozygous variant genotype at the genomic coordinate 502 (e.g., "L(0 / 1)@chr5:4"), and iii) a third genotype probability 514 that the initial genotype call 505 is a homozygous variant genotype at the genomic coordinate 502 (e.g., "L(1 / 1)@chr5:4").

[0166] As mentioned, the call integration system 106 generates genotype probabilities 508 to predict whether a SNP occurs at the genomic coordinates 502. However, in order to predict whether an indel occurs at the genomic coordinates, the call integration system 106 generates a different set of machine learning predictions. Specifically, the call integration system 106 generates variant call classifications that indicate the presence (or absence) of an indel (or a multi-allelic SNP or another variant type other than a biallelic SNP) at the genomic coordinates of the sample sequence.

[0167] like Figure 5B As shown, the detection integration system 106 uses the genotype detection integration machine learning model 520 to generate variant detection classification 522. In detail, the detection integration system 106 uses the genotype detection integration machine learning model 520 to generate variant detection classification 522 based on sequencing metrics 518 and initial genotype detection 519 associated with genomic coordinates 516. In fact, similar to the above discussion on generating genotype probabilities for biallelic SNPs, the detection integration system 106 also determines the sequencing metrics 518 associated with the genomic coordinates 516, which include sequencing metrics based on reads, sequencing metrics generated by the detection model, and sequencing metrics from external sources. For example, the detection integration system 106 uses the detection generation model 517 to analyze a subset of the sequencing metrics 518 (e.g., sequencing metrics based on reads and / or sequencing metrics from external sources) for determining the initial genotype detection 519 (e.g., indicating a specific genotype or variant at the genomic coordinates 516). In some cases, call generation model 517 further generates a subset of sequencing metrics 518 associated with genomic coordinates 516 (eg, call model-generated sequencing metrics).

[0168] When generating a variant call classification 522 for the genomic coordinate 516, the call integration system 106 utilizes the genotype call integration machine learning model 520. Specifically, the call integration system 106 utilizes the genotype call integration machine learning model 520 to generate: i) a first true positive variant probability 524, which indicates the possibility that the initial genotype call 519 (or initial VCF file) from the first type of nucleotide read (e.g., SBS read) is a true positive at the genomic coordinate 516, ii) a second true positive variant probability 526, which indicates the possibility that the initial genotype call 519 (or initial VCF file) from the second type of nucleotide read (e.g., assembled nucleotide read) is a true positive at the genomic coordinate 516, iii) a first joint 516. The variant call classifications 522 include, for example, a first zygosity error probability 528 indicating the likelihood that the initial genotype call 519 (or initial VCF file) from the second type of nucleotide reads exhibits a genotype-zygosity error at genomic coordinate 516, iv) a second zygosity error probability 530 indicating the likelihood that the initial genotype call 519 (or initial VCF file) from the second type of nucleotide reads exhibits a genotype-zygosity error at genomic coordinate 516, and v) a reference probability 532 indicating the likelihood that the initial genotype call 519 at genomic coordinate 516 is a homozygous reference genotype (or a false positive). In some cases, the variant call classifications 522 are mutually exclusive.

[0169] As shown, the first true positive variant probability 524 is represented by "TP_s". The symbol "TP_s" represents the probability that the input (x) is a true positive variant in the first variant call file (e.g., SBS variant call file), where "TP_s" can be expressed as P(tp_s|x)), and "s" represents a first type of nucleotide read, in particular a "short read" or SBS read. In addition, the second true positive variant probability 526 is represented by "TP_l". The symbol "~TP_s&TP_l" represents the probability that the input (x) is not a true positive in the first variant call file (e.g., SBS variant call file) and is a true positive in the second variant call file (e.g., assembled nucleotide read variant call file), where "~TP_s&TP_l" can be expressed as P(~tp_s&p_1|x)), and where "l" represents a "long read" or assembled nucleotide read.

[0170] In contrast, the first zygosity error probability 528 is represented by "HH_s". The notation "~TP_s & TP_l & HH_s" represents the probability that the input (x) is not a true positive in the first variant call file (e.g., SBS variant call file), is not a true positive in the second variant call file (e.g., assembled nucleotide read variant call file), and is a het / hom error in the first variant call file (e.g., SBS variant call file). Additionally, the second zygosity error probability 530 is represented by "HH_l". The symbols "~TP_s&~TP_l&~HH_s&HH_l" represent the probability that the input (x) is not a true positive in the first variant call file (e.g., SBS variant call file), is not a true positive in the second variant call file (e.g., assembled nucleotide read variant call file), is not a het-hom error in the first variant call file (e.g., SBS variant call file), and is a het-hom error in the second variant call file (e.g., assembled nucleotide read variant call file). In addition, a reference probability 532 is represented by "FP", which indicates the probability that the input (x) is a false positive and can be expressed by the formula P(fp|x)).

[0171] To elaborate on the first zygosity error probability 528 and the second zygosity error probability 530, the call integration system 106 determines the probability that the predicted genotype at the genomic coordinate 516 (e.g., the initial genotype call for the different read types) is an incorrect genotype (e.g., a genotype incorrectly identified by the call generation model 517) or includes an incorrect allele. In detail, in some cases, the call integration system 106 determines the probability that there is a zygosity error (e.g., a het / hom error) at the genomic coordinate 516 (e.g., where the alternative base is correct but the genotype is incorrect) based on the first type of nucleotide reads or the second type of nucleotide reads, or the probability that the genotype call represents a completely incorrect genotype or an error allele in the initial genotype call 519. For example, when determining the probability of a zygosity error, the call integration system 106 determines the probability that the alternative base call represented as "1" is correct but the genotype is incorrect, such as the probability of incorrectly determining a 0 / 1 genotype call (e.g., A / T) instead of a correct 1 / 1 genotype call (e.g., T / T) (or vice versa in which the correct genotype call is 0 / 1).

[0172] The call integration system 106 can correct the inaccuracies of existing sequencing systems, where incorrect calls are often insertions and deletions, by determining the first zygosity error probability 528 and the second zygosity error probability 530. Specifically, the call integration system 106 can more accurately generate genotype calls corresponding to the genomic coordinates of the insertion and deletion, in which case the existing sequencing system would determine that the genotype call represents an incorrect genotype, which represents an incorrect allele resulting from a long insertion or deletion sequence.

[0173] like Figure 5B As further illustrated in , the call integration system 106 utilizes the genotype call integration machine learning model 520 to generate a first true positive variant probability 524 and a second true positive variant probability 526. Specifically, the call integration system 106 generates a first true positive variant probability 524 based on a first type of nucleotide read (e.g., SBS read), and generates a second true positive variant probability 526 based on a second type of nucleotide read (e.g., assembled nucleotide read). In some cases, the true positive variant probability indicates the probability of a correct variant call genotype at the genomic coordinate 516. For example, the call integration system 106 generates a probability that the initial genotype call 519 for the genomic coordinate 516 is correct as determined by the call generation model 517.

[0174] Continue to Figure 5C In some embodiments, the call integration system 106 updates one or more data fields or variant call file fields ("VCF" fields) associated with the variant call file using the genotype probabilities 508 and / or the variant call classifications 522. For example, the call integration system 106 generates a merged SNP variant call file 536 based on the genotype probabilities 508 and the variant call classifications 522. In fact, in some cases, the call integration system 106 generates a single merged variant call file that combines data from the genotype probabilities 508 for SNPs and from the variant call classifications 522 for indels.

[0175] As shown, the call integration system 106 generates updated VCF fields 534 that indicate or correspond to updated sequencing metrics for the output genotype calls. Specifically, the call integration system 106 generates one set of updated VCF fields for genotype probability 508 and another set of updated VCF fields for variant call classification 522. For illustrative purposes, Figure 5CSeveral example fields within updated VCF fields 534 are shown, without separately depicting one set of updated VCF fields for genotype probabilities 508 and another set of updated VCF fields for variant call classifications 522. In some cases, the call integration system 106 modifies or updates only certain VCF fields based on genotype probabilities 508 and / or variant call classifications 522 and does not update other fields.

[0176] In other cases, the call integration system 106 does not update the VCF fields. For example, when generating a genotype call, the call integration system 106 does not update certain fields, such as the genotype (GT) field, based on the genotype probability 508 and / or the variant call classification 522. In fact, in some cases, the call integration system 106 does not modify or update the GT field because there may not be enough information to determine a new or updated genotype at the genomic coordinate.

[0177] To illustrate one embodiment, Figure 5C Depicted is a call integration system 106 generating an updated VCF field 534 for a 1 / 2 genotype (GT), wherein cytosine represents a reference base at a genomic coordinate corresponding to an allele of a reference genome (shown as "Ref:C"), adenine represents a first alternative base at a genomic coordinate of a different allele ("Alt1:A"), and thymine represents a second alternative base at a genomic coordinate of yet another different allele ("Alt 2:T"). Figure 5C Only examples of possible reference bases and possible alternative bases at genomic coordinates are depicted. The call integration system 106 can generate genotype probabilities 508 and variant call classifications 522 to modify corresponding metrics for various other reference bases and alternative bases at genomic coordinates in the VCF field.

[0178] like Figure 5CAs further illustrated in , the detection integration system 106 generates an updated base detection quality (QUAL) field. More specifically, the detection integration system 106 modifies or updates the base detection quality metric based on the genotype probability 508 and / or the variant detection classification 522 to indicate the accuracy of the genotype detection. As shown in the figure, the updated base detection quality field indicates the QUAL score 48 of the variant at the corresponding genomic coordinate. In this example, the updated base detection quality metric (e.g., QUAL score 48) represents the score of any type of variant at the corresponding genomic coordinate. In addition, the detection integration system 106 generates a modified or updated genotype quality (GQ) field. For example, based on the variant detection classification 522, the detection integration system 106 generates a modified or updated genotype quality metric, which modifies or updates the genotype quality metric indicating the possibility or probability that the predicted genotype at the genomic coordinate is correct. As shown, for example, the updated genotype quality field indicates a genotype quality metric for genotype calls having a heterozygous genotype (eg, a GQ score of 4 for genotype 1 / 2 at a multi-allelic genomic coordinate).

[0179] In one or more embodiments, the detection integration system 106 further generates or updates the genotype probability field, and (in some cases) uses the genotype probability field to rank alleles. In detail, the detection integration system 106 generates the GT field of update by sorting the candidate genotype detection at the genome coordinates according to the corresponding probability belonging to the multi-allelic genome coordinates. For example, the detection integration system 106 determines the probability associated with a plurality of genotypes, wherein each diploid genotype is composed of a pair of alleles. For example, the detection integration system 106 determines the relative probability of belonging to the genome coordinates associated with a plurality of alleles (for example, from a reference genome, the first alternative allele and the second alternative allele).

[0180] In some embodiments, the call integration system 106 also or alternatively generates a metric for a PHRED scaled likelihood (PL) field as part of the updated VCF fields. For example, the call integration system 106 generates a metric for a PL field that can indicate a genotype, such as a homozygous reference genotype, a heterozygous genotype, and a homozygous alternative genotype (e.g., having PL field designations 9 / 0 / 3, respectively).

[0181] In one or more embodiments, the call integration system 106 generates allele-specific probabilities or possibilities based on the relative probabilities of genotype calls corresponding to alleles from the call generation model and any other (non-reference) genotypes identified by the genotype call integration machine learning model. For example, in some embodiments, the call integration system 106 indicates the relative probability score of each allele of the corresponding genotype call in the PL field corresponding to the normalized PHRED scaled likelihood of the indicated genotype and / or the genotype likelihood (GP) field indicating the logarithmic scaled posterior genotype probability (e.g., log10 scaling) of the data (e.g., sequencing metrics) given the genotype called.

[0182] As a motivation for modifying certain VCF fields of SNPs, in some cases, the call integration system 106 utilizes a genotype call integration machine learning model to generate genotype probabilities 508 (whose probabilities sum to 1). Specifically, the genotype call integration machine learning model may generate a first genotype probability 510 as 0.1, a second genotype probability 512 as 0.2, and a third genotype probability 514 as 0.7. In such an example, based on the genotype probabilities 508, the call integration system 106 generates an updated genotype probability field by updating the GT field, the GP field, and the PL field using a combination of information from the genotype call integration machine learning model and the call generation model.

[0183] like Figure 5C As further illustrated in , the detection integration system 106 updates the PL field for different genotypes (GT). According to the normalized scale of the PL score, a relatively low score for a genotype (e.g., PL 0) indicates a relatively high probability that the genotype exists at the genomic coordinate; and a relatively high score for a genotype (e.g., PL 101) indicates a relatively low probability that the genotype exists at the genomic coordinate. For example, the detection integration system 106 determines that the PL score for the 0 / 0 genotype is 111, the PL score for the 0 / 1 genotype is 52, and the PL score for the 1 / 1 genotype is 52. Therefore, in Figure 5C , a PL score of 52 indicates the genotype with the highest likelihood or the selected genotype (eg, 0 / 1 and 1 / 1 genotypes), and a PL score of 111 indicates the lowest likelihood (eg, 0 / 0 genotype).

[0184] In some cases, the call integration system 106 generates the updated genotype probability field as a ranking of multiple alleles identified via the call generation model (without using the genotype call integration machine learning model). In other cases, the call integration system 106 uses a specialized version of the genotype call integration machine learning model that is trained to generate an updated genotype probability field based on the genotype probability 508 and / or variant call classification 522.

[0185] like Figure 5C As further illustrated in , the call integration system 106 generates or updates a variant call file, such as a merged SNP variant call file 536. For example, the call integration system 106 generates a variant call file based on the updated VCF fields 534 corresponding to the genotype probability 508 and the variant call classification 522, respectively. Therefore, the call integration system 106 generates a merged SNP variant call file 536 for SNP genotype calls based on the genotype probability 508 and / or the variant call classification 522. In fact, in some embodiments, the call integration system 106 generates a merged variant call file that merges the data of SNPs and indels from both the genotype probability 508 and the variant call classification 522.

[0186] like Figure 5C As indicated, the call integration system 106 may generate a merged SNP variant call file 536 to include updated VCF fields 534, which include base call quality metrics, genotype quality metrics, and / or updated genotype probability fields. For example, the call integration system 106 selects VCF fields from the initial genotype calls generated by the call generation model (such as initial genotype calls for SBS reads and initial genotype calls for assembled nucleotide reads) to include in the merged variant call file. However, in some embodiments, the call integration system 106 does not select fields, but instead generates new VCF fields for the merged variant call file by processing the genotype probability 508 and the variant call classification 522 using the genotype call integration machine learning model.

[0187] As mentioned, in some cases, the detection integration system 106 only updates certain fields, while other fields such as the genotype (GT) field remain unchanged. For example, the detection integration system 106 updates the genotype quality field and the base detection quality field. For other data fields such as the normalized PHRED scaled possibility (PL) and the posterior genotype probability (GP) of the genotype, the detection integration system 106: (i) maintains the field as it is, (ii) removes the field, or (iii) updates the field for reflecting the detected genotype and the GQ of category 0 output 0 / 0. In some cases, the detection integration system 106 maintains the relative probability of other genotypes relative to the detected genotype to ensure consistent updates and the detected genotype is the highest. In certain embodiments, by only updating the values ​​of 0 / 0 and 1 / 2, the detection integration system 106 maintains the distance between other genotypes and the detected genotype. By updating only certain fields, the call integration system can generate (merged) variant call files more efficiently, without having to regenerate an entirely new variant call file (as some existing systems do) and / or update every field (even those that have not changed due to new predictions).

[0188] Within (or as a result of generating) the merged variant call file, the call integration system 106 may include or update one or more output genotype calls (e.g., variant calls) associated with the genomic coordinates, as determined based on the updated VCF field 534. In practice, to generate the output genotype calls, the call integration system 106 may predict the nucleobases based on the candidate alleles at the genomic coordinates (e.g., based on their corresponding probabilities and metrics indicated by the merged variant call file). Thus, the call integration system 106 may generate output SNPs and / or output indel calls based on the merged SNP variant call file 536.

[0189] Because the detection integration system 106 generates genotype calls based on multiple read types in a single pipeline (e.g., combining data from each type of read), there are some situations where different types of nucleotide reads conflict. In fact, in some cases, the alternative reads of the first type of nucleotide reads (e.g., SBS reads) and the alternative reads of the second type of nucleotide reads (e.g., assembled nucleotide reads) may be inconsistent, where different read types indicate different nucleotide bases. In such cases, the detection integration system 106 can utilize a machine learning model that is trained to determine which read data is more accurate between different read types (e.g., by determining which replacement to select between SBS reads and assembled nucleotide reads). In some embodiments, the detection integration system 106 resolves conflicts or disagreements between different read types by automatically selecting the replacement indicated by the SBS rather than the replacement indicated by the assembled nucleotide reads (or other read types).

[0190] As described above, in certain embodiments, the detection integration system 106 trains or adjusts the genotype detection integration machine learning model by learning model parameters (such as weights and biases for generating accurate genotype probabilities or accurate variant detection classifications). Specifically, the detection integration system 106 uses an iterative training process to fit or train the genotype detection integration machine learning model by adjusting or adding decision trees or learning parameters that result in generating genotype probabilities (for SNPs) and / or variant detection classifications (for insertions and deletions). Figure 6 A call integration system 106 for training a genotype call integration machine learning model according to one or more embodiments is illustrated. Figure 6 Different instances of the genotype calling ensemble machine learning model are depicted to concisely illustrate the training process, but in some embodiments, the call integration system 106 trains and adjusts model parameters of one instance or version of the genotype calling ensemble machine learning model and another instance or version of the genotype calling ensemble machine learning model 608 separately from each other. Figure 6 As depicted in , the detection integration system 106 separately trains the genotype detection integration machine learning model 606 (e.g., SNP-specific model) and the genotype detection integration machine learning model 608 (e.g., insertion-deletion-specific model) as different machine learning models based on different benchmark truth data. Although trained as different machine learning models, in some cases, the genotype detection integration machine learning model 606 and the genotype detection integration machine learning model 608 each include the same type of machine learning model (e.g., gradient boosting decision tree, deep learning transformer).

[0191] like Figure 6As illustrated in , the detection integration system 106 trains an instance of the genotype detection integrated machine learning model 606 to generate a genotype probability for a SNP, and trains another instance of the genotype detection integrated machine learning model 608 to generate a variant detection classification for insertion and deletion. Specifically, the detection integration system 106 accesses the sample sequencing metrics 604 from the database 602 for use as training data. For example, the detection integration system 106 accesses the sample sequencing metrics 604, which include metrics based on sample reads, sequencing metrics from external sources of the sample, and sequencing metrics generated by the sample detection model. In certain embodiments, the sample sequencing metrics 604 can be determined, generated, or derived from a plurality of different genomic samples analyzed or processed by different sequencing devices. In fact, the detection integration system 106 can use sample sequencing metrics 604 with different variability dimensions to train the genotype detection integrated machine learning model 606 and / or the genotype detection integrated machine learning model 608. Specifically, the sample sequencing metrics 604 can vary in terms of the coverage or amount of sequencing performed on the sample to obtain the sequencing metrics. The sample sequencing metrics 604 may also (or alternatively) vary in the library preparation method, the sequencing equipment used to obtain the sample sequencing metrics 604, the quality of the sequencing run (e.g., Q30, error rate, and / or %PF passing a percentage filter).

[0192] In some cases, the sample sequencing metric 604 has a corresponding reference truth variant call file (e.g., as part of the reference truth data 620) associated therewith (e.g., stored in the database 602), wherein the reference truth variant call file indicates the actual VCF fields of the actual genotype calls generated by the sample sequencing metric 604. For example, the call integration system 106 utilizes the sample sequencing metric 604 and a reference truth variant call file (e.g., as part of the reference truth 620) from a training data set generated by the U.S. Food and Drug Administration (referred to as the PrecisionFDA data set). In some cases, the sample sequencing metric 604 includes a subset of the sample sequencing metric for each genotype call in the reference truth variant call file. The reference truth variant call file may have reference truth genotype calls corresponding to the sample sequencing metric.

[0193] As mentioned, the call integration system 106 trains a genotype call integration machine learning model 606 for SNP genotype calling. To train the genotype call integration machine learning model 606, the call integration system 106 inputs the sample sequencing metrics 604 and the sample genotype calls 603 (e.g., initial genotype calls generated by a call generation model from the sample sequencing metrics 604) into the genotype call integration machine learning model 606. Then, the genotype call integration machine learning model 606 generates a predicted genotype probability 610 based on the sample sequencing metrics 604. For example, as described above, the genotype call integration machine learning model 606 generates a predicted first genotype probability, a predicted second genotype probability, and a predicted third genotype probability.

[0194] As part of training the genotype call ensemble machine learning model 608 for indels, the call integration system 106 inputs the sample sequencing metrics 604 and the sample genotype calls 603 into the genotype call ensemble machine learning model 608. Then, the genotype call ensemble machine learning model 608 generates a predicted variant call classification 612 based on the sample sequencing metrics 604. Specifically, in some embodiments, the genotype call ensemble machine learning model 608 generates a set of five predicted variant call classifications, including a first true positive variant probability, a second true positive variant probability, a first zygosity error probability, a second true positive zygosity error probability, and a reference probability, as described above.

[0195] Based on the predicted genotype probabilities 610 and / or the predicted variant call classifications 612, the call integration system 106 generates a modified variant call file 614. For example, the call integration system 106 generates a modified variant call file based on the predicted genotype probabilities 610 for training the genotype call integration machine learning model 606. Additionally or alternatively, the call integration system 106 generates a modified variant call file based on the predicted variant call classifications 612 for training the genotype call integration machine learning model 608.

[0196] like Figure 6As further illustrated in , the call integration system 106 performs a comparison 616. Specifically, the call integration system 106 performs a comparison 616 to compare (i) the predicted genotype probability 610 with the benchmark truth data 620 (e.g., the benchmark truth genotype probability) and / or (ii) the predicted variant call classification 612 with the benchmark truth data 620 (e.g., the benchmark truth variant call classification). In some embodiments, the call integration system 106 uses a loss function 618 to perform the comparison 616. For example, the call integration system 106 uses a cross entropy loss function to compare the predicted genotype probability 610 with the benchmark truth genotype probability and / or the predicted variant call classification 612 with the benchmark truth variant call classification (e.g., to determine a measure of error or loss therebetween). In the case where the genotype call ensemble machine learning model 606 or 608 is an ensemble of gradient boosted trees, the call ensemble system 106 utilizes a mean squared error loss function (e.g., for regression) and / or a logarithmic loss function (e.g., for classification) as the loss function 618.

[0197] In contrast, in embodiments where the genotype call ensemble machine learning model 606 is a neural network, the call ensemble system 106 may utilize a cross entropy loss function, an L1 loss function, or a mean square error loss function as the loss function 618. For example, the call ensemble system 106 utilizes the loss function 618 to determine the difference between the predicted genotype probabilities 610 and the baseline true genotype probabilities of the baseline true value data 620 and / or the predicted variant call classifications 612 and the baseline true value variant call classifications of the baseline true value data 620.

[0198] In some embodiments, the call integration system 106 may utilize (i) a call generation model to generate initial genotype calls, and (ii) a genotype call integration machine learning model 606 or 608 to modify data fields of a variant call file corresponding to the initial genotype calls to generate new predicted genotype calls. The call integration system 106 outputs such modified or recalibrated values ​​as part of a modified variant call file 614. For example, the call integration system 106 determines recalibrated values ​​for metrics within the modified variant call file 614, including a call quality metric (QUAL), a genotype metric (GT), and a genotype quality metric (GQ).

[0199] like Figure 6As further illustrated in , the call integration system 106 performs model fitting 622. Specifically, the call integration system 106 fits the genotype call integration machine learning model 606 or 608 based on the comparison 616. For example, the call integration system 106 performs modifications or adjustments to the parameters (e.g., weights and biases) of the genotype call integration machine learning model 606 or 608 to reduce the measure of loss from the loss function 618 and uses the adjusted parameters on subsequent training iterations.

[0200] For gradient boosting trees, for example, the call integration system 106 trains the genotype call integration machine learning model 606 or 608 on the error gradient determined by the loss function 618. For example, the call integration system 106 solves a (e.g., infinite-dimensional) convex optimization problem while regularizing the objective to avoid overfitting. In some specific implementations, the call integration system 106 scales the gradient to emphasize corrections for underrepresented classes (e.g., where there are significantly more true positive variant calls than false positive variant calls).

[0201] In some embodiments, as part of solving the optimization problem, the call integration system 106 adds new weak learners (e.g., new boosted trees) to the genotype call integration machine learning model 606 or 608 for each successive training iteration. For example, the call integration system 106 finds a feature (e.g., a sequencing metric) that minimizes the loss from the loss function 618 and adds the feature to the tree of the current iteration or begins building a new tree using the feature.

[0202] In addition to or in lieu of gradient boosting decision trees, the call integration system 106 trains logistic regression to learn parameters for generating genotype calls. To avoid overfitting, the call integration system 106 is further regularized based on hyperparameters such as learning rate, stochastic gradient boosting, number of trees, tree depth, complexity penalty, and L1 / L2 regularization.

[0203] In an embodiment where the genotype detection integration machine learning model 606 or 608 is a neural network, the detection integration system 106 reduces the measure of the loss of the loss function 618 by modifying the internal parameters (e.g., weights) of the genotype detection integration machine learning model 606 or 608, thereby performing model fitting 622. In practice, the detection integration system 106 modifies how the genotype detection integration machine learning model 606 or 608 analyzes and passes data between layers and neurons by modifying the internal network parameters. Therefore, through multiple iterations, the detection integration system 106 improves the accuracy of the genotype detection integration machine learning model 606 or 608.

[0204] In fact, in some cases, the detection integration system 106 repeats Figure 6The training process shown is repeated for multiple iterations. For example, the detection integration system 106 repeats iterative training by selecting a new sequencing metric set and a corresponding benchmark variant detection file for sample genotype detection. The detection integration system 106 further generates a new set of predicted genotype probabilities and / or variant detection classifications and a new modified variant detection file in each iteration. As described above, the detection integration system 106 also compares the genotype detection and / or data fields from the modified variant detection files with the detection and / or data fields from the corresponding benchmark true value variant detection files at each iteration. The detection integration system 106 further performs model fitting for each iteration. The detection integration system 106 repeats the process until the genotype detection integrated machine learning model 606 or 608 generates a predicted genotype probability or variant detection classification, and these predicted genotype probabilities or variant detection classifications produce genotype detections or variant detection files that meet the threshold measurement of loss.

[0205] In some cases, the detection integration system 106 uses the validation data set to determine when training is complete. For example, the detection integration system 106 determines the loss of the validation data set (e.g., by comparing the validation data with the predicted genotype probability 610 and / or the predicted variant detection classification 612). Based on determining that the loss value associated with the validation data set has not decreased (by a threshold amount) for at least a threshold number of iterations (e.g., 10 iterations), the detection integration system 106 may determine that training is complete. In some embodiments, the detection integration system 106 may perform training for a threshold number of iterations (e.g., 400 iterations), after which the detection integration system 106 determines that training is complete.

[0206] although Figure 6 Not illustrated, but in certain embodiments, the call integration system 106 trains and adjusts model parameters for a single genotype call integration machine learning model to generate different outputs (e.g., genotype probabilities and variant call classifications) in different training iterations or training epochs. For example, the call integration system 106 (i) performs a set of training iterations to train and adjust model parameters for the genotype call integration machine learning to generate genotype probabilities, and (ii) performs another set of training iterations to train and adjust the same genotype call integration machine learning model to generate variant call classifications. However, because two different genotype call integration machine learning models (e.g., a SNP-specific genotype call integration machine learning model and an insertion and deletion-specific genotype call integration machine learning model) perform better in recovering false positive variants and false negative variants, Figure 6 A separately trained genotype calling ensemble machine learning model 606 and / or a genotype calling ensemble machine learning model 608 are depicted.

[0207] As mentioned, in certain described embodiments, the call integration system 106 utilizes the genotype call integration machine learning model and the call generation model to generate genotype calls. Specifically, the call integration system 106 utilizes the output of the genotype call integration machine learning model to modify data fields corresponding to a variant call file that includes genotype calls initially generated by the call generation model. Figure 7 Illustrated is a call integration system 106 that generates genotype calls based on the output of a genotype call integration machine learning model and a call generation model and modifies fields of a variant call file that includes genotype calls and reported metrics according to one or more embodiments.

[0208] like Figure 7 As illustrated in , the detection integration system 106 accesses the sequencing information database 702, the reference sequence 704, and the sequence data 708 extrapolated from one or more nucleotide reads (e.g., the first type of nucleotide reads and / or the second type of nucleotide reads). In fact, the detection integration system 106 performs a sequencing metric extraction 714 to extract or reengineer the sequencing metrics as described above. For example, the detection integration system 106 generates sequencing metrics based on reads, sequencing metrics from external sources, and sequencing metrics generated by the detection model. In some cases, the detection integration system 106 uses the mapping and alignment component 710 of the detection generation model 724 to determine the mapping and alignment sequencing metrics as described above. In addition, the detection integration system 106 uses the variant detector component 712 of the detection generation model 724 to generate the variant detection metrics as described above. In addition, the detection integration system 106 determines the sequencing metrics based on the reads and the sequencing metrics from external sources (e.g., based on the sequencing information database 702 and / or the reference sequence 704).

[0209] like Figure 7 As further illustrated in , the detection integration system 106 generates a genotype probability 716 and / or a variant detection classification 718. By analyzing the sequencing metrics, the first genotype detection 700a corresponding to the first type of nucleotide reads, and the second genotype detection 700b corresponding to the second type of nucleotide reads, the detection integration system 106 uses the genotype detection integration machine learning model 706a to generate the genotype probability 716 of the SNP, as described herein. In addition, by analyzing the sequencing metrics, the first genotype detection 700a corresponding to the first type of nucleotide reads, and the second genotype detection 700b corresponding to the second type of nucleotide reads, the detection integration system 106 uses the genotype detection integration machine learning model 706b to generate a variant detection classification 718 for insertion and deletion, as described herein. As described above, the first genotype detection 700a corresponding to the first type of nucleotide reads and the second genotype detection 700b corresponding to the second type of nucleotide reads may come from different read type pipelines.

[0210] In some cases, the genotype call ensemble machine learning model 706a or 706b is an ensemble of gradient boosted trees that process sequencing metrics to generate genotype probabilities 716 or variant call classifications 718. For example, the genotype call ensemble machine learning model 706a or 706b includes a series of weak learners, such as non-linear decision trees trained in logistic regression to generate genotype probabilities 716 or variant call classifications 718. In some cases, the genotype call ensemble machine learning model 706a or 706b includes various intra-tree metrics that define how to process sequencing metrics to generate corresponding outputs based on the above training.

[0211] As mentioned above, in some embodiments, the call integration system 106 can utilize the genotype call integration machine learning models 706a and 706b together. For example, the call integration system 106 utilizes the genotype call integration machine learning models 706a and 706b to generate genotype probabilities 716 and variant call classifications 718, respectively. For example, the call integration system 106 utilizes two (or more) different genotype call integration machine learning models in parallel, each genotype call integration machine learning model is trained with a different random seed (e.g., processes data differently for different biases) and / or is trained on different training data for different types of variants, resulting in different prediction outputs.

[0212] In some embodiments, the detection integration system 106 further generates a combined prediction set from the outputs of different genotype detection integration machine learning models 706a and 706b. For example, the detection integration system 106 combines (e.g., averages or totals) metrics from genotype probability 716 and variant detection classification 718. In some embodiments, the detection integration system 106 determines the mean between predictions from different models and renormalizes the mean. In other embodiments, the detection integration system 106 learns linear weights and adapts weights to minimize total error or loss. In other embodiments, the detection integration system 106 weights genotype probability and / or variant detection classification for the corresponding genotype detection integration machine learning model based on the inverse of the average error across models.

[0213] In one or more specific implementations, the detection integration system 106 further utilizes a meta-model following the genotype detection integration machine learning models 706a and 706b. For example, the detection integration system 106 generates a genotype probability 716 (e.g., genotype probability 508) and a variant detection classification 718 (e.g., variant detection classification 522) as described above, and combines them using a classification combiner machine learning model. Specifically, the detection integration system 106 can combine the genotype probability and variant detection classification generated from each genotype detection integration machine learning model by selecting the weights applied to the variant detection classification generated by each genotype detection integration machine learning model. In fact, in some cases, the detection integration system 106 trains the classification combiner machine learning model to determine, select or predict the corresponding weights of the genotype detection integration machine learning model to produce the highest accuracy or minimum loss.

[0214] As an example of generating genotype probabilities 716 and / or variant call classifications 718, in some embodiments, the call integration system 106 uses statistics to summarize the mapping quality distributions of reference support reads and alternative support reads (e.g., for comparing mapping quality distribution metrics). The call integration system 106 can determine and use the mean of the MAPQ to support reads of alternative alleles from SBS reads and from assembled nucleotide reads. In these or other embodiments, the genotype call integration machine learning model 706a or 706b learns from the data that when the MAPQ of the alternative allele (represented by the SBS read or assembled nucleotide read) is low and the depth metric is high relative to other MAPQs and depth metrics in the distribution, the resulting genotype call is more likely to be a false positive. In fact, as the probability of a false positive increases, the MAPQ metric will likely decrease.

[0215] For another example, in some cases, the detection integration system 106 compares the mapping quality (e.g., MAPQ) associated with the SBS reads and / or assembled nucleotide reads with a mapping quality threshold. For example, the detection integration system 106 utilizes a mapping quality threshold, such as a threshold difference between the best alignment score and the second-best alignment score. When it is determined that one or more of the mapping qualities for different read types do not meet the threshold, the detection integration system 106 adjusts one or more of the genotype probability 716 or the variant call classification 718 accordingly (e.g., selects a read with a higher MAPQ).

[0216] Additionally (or alternatively), the call integration system 106 can determine how to best fit the data by utilizing an accumulation of statistical analyses of complex functions (depending on the architecture of the genotype call integration machine learning model 706a or 706b) to determine the genotype probability 716 and / or variant call classification 718. For example, as described above, the call integration system 106 trains the genotype call integration machine learning model 706a or 706b to minimize the loss generated from multiple (different types of) sequencing metrics to determine the weights and biases that best fit the data (e.g., resulting in reduced or minimized loss).

[0217] like Figure 7 As further illustrated in , in addition to generating genotype probabilities 716 and variant call classifications 718, the call integration system 106 also performs data field generation 720. More specifically, the call integration system 106 generates data fields for one or more variant call files. In some cases, the call integration system 106 generates a first variant call file including a first genotype call 700a and further generates a second variant call file including a second genotype call 700b. As mentioned, the call integration system 106 can use the first genotype call 700a and / or the second genotype call 700b to generate predictions, such as genotype probabilities 716 and variant call classifications 718. As further shown, the call integration system 106 can use data field generation 720 to generate a merged variant call file 722 (e.g., by combining all of the first variant call file and the second variant call file or selecting a portion of these files) to indicate the output genotype call. To generate the merged variant call file 722 , the call integration system 106 utilizes the variant caller component 712 of the call generation model 724 and modifies or maintains the values ​​of such data fields based on the genotype probability 716 and / or the variant call classification 718 .

[0218] For example, the detection integration system 106 modifies various metrics, such as quality metrics, mapping metrics, or other metrics associated with genotype detection. As mentioned, in some cases, the detection integration system 106 selects metrics associated with the first type of nucleotide reads or the second type of nucleotide reads and / or associated with the genotype probability 716 of the SNP and / or the variant detection classification 718 of the insertion and deletion. In other cases, the detection integration system 106 generates new metrics from data generated by the detection generation model 724 and / or the genotype detection integration machine learning model 706a or 706b. In certain embodiments, genotype detection is represented or defined by a merged variant detection file 722, which includes metrics corresponding to data fields, such as detection quality metrics corresponding to the detection quality field, genotype metrics corresponding to the genotype field, and genotype quality metrics corresponding to the genotype quality field.

[0219] In certain embodiments, the call integration system 106 generates (data fields for) genotype calls using the variant caller component 712 and the genotype probability 716 and / or the variant call classification 718. For example, the call integration system 106 generates data fields for various metrics of genotype calls for inclusion in the merged variant call file 722 and using the variant caller component 712, such as the nucleotide included in the call, the quality of the call (QUAL), the genotype (GT), the genotype quality (GQ), one or more normalized PHRED scaled likelihoods (PL), and / or the genotype probability (GP).

[0220] In one or more embodiments, the call integration system 106 recalibrates or modifies genotype calls (or generates new genotype calls) using the genotype probabilities 716 from the genotype call integration machine learning model 706a and / or the variant call classifications 718 from the genotype call integration machine learning model 706b. As described, the call integration system 106 modifies the genotype calls by modifying or recalibrating the data fields of one or more metrics associated with the genotype calls (e.g., as included in the merged variant call file 722).

[0221] For example, in order to update or recalibrate the detection quality metric (QUAL) associated with genotype detection, the detection integration system 106 determines how each of the genotype probability 716 and / or variant detection classification 718 impacts or affects the base detection quality metric. For example, the detection integration system 106 determines that a high probability of genotype error results in a lower overall genotype quality and a possibly different overall detection quality. For another example, the detection integration system 106 determines that a high probability of a false positive variant results in a lower overall detection quality. For another example, the detection integration system 106 determines that a high probability of a true positive variant results in a higher overall (variant) detection quality. The detection integration system 106 updates the genotype and the genotype quality and detection quality associated with the genotype detection accordingly.

[0222] In one or more specific implementations, the detection integration system 106 generates a combination (e.g., a weighted combination or average) of genotype probabilities 716 and / or variant detection classifications 718 to recalibrate the detection quality metric. Specifically, the detection integration system 106 weights the genotype probabilities 716 and / or variant detection classifications 718 according to their corresponding impact on (variant) detection quality. In some cases, the detection integration system 106 evenly weights each genotype probability or variant detection classification, while in other cases, the detection integration system 106 determines different weights for each genotype probability or variant detection classification. In any case, the detection integration system 106 determines a weighted combination or weighted average of genotype probabilities 716 and variant detection classifications 718 to recalibrate (increase or decrease) the detection quality metric of genotype detection (e.g., initial variant detection).

[0223] To update or recalibrate the genotype metric associated with the genotype call (e.g., within the GT field of the merged variant call file 722), the call integration system 106 utilizes one or more of the genotype probabilities 716 and / or the variant call classifications 718. For example, the call integration system 106 compares the various component predictions of each to determine which of the genotype probabilities 716 or variant call classifications 718 has the highest probability. In some cases, the call integration system 106 recalibrates the genotype metric using the genotype probabilities and / or variant call classifications with the highest probability (e.g., from 0, such as corresponding to the reference base, to 1, such as corresponding to the first alternative supporting read).

[0224] In order to update or recalibrate the genotype quality metric associated with the genotype call (e.g., within the GQ field of the merged variant call file 722), the call integration system 106 utilizes one or more of the genotype probability 716 and / or the variant call classification 718. More specifically, the call integration system 106 determines how each of the genotype probability 716 and / or the variant call classification 718 affects the genotype quality metric. The call integration system 106 recalibrates the genotype quality metric accordingly (e.g., by increasing or decreasing the quality score between 0 and 10 or between 0 and 100 or on some other scale). For example, the call integration system 106 determines that a higher genotype error probability (usually) indicates a lower genotype quality metric, and the call integration system 106 reduces the metric accordingly.

[0225] In some cases, the detection integration system 106 determines a combination (e.g., a weighted combination or a weighted average) of the genotype probability 716 and / or the variant detection classification 718 to modify the genotype quality metric. For example, the detection integration system 106 determines the combined effect of the genotype probability 716 and / or the variant detection classification 718 on the genotype quality metric. As another example, the detection integration system 106 determines the individual impact of each component prediction of the genotype probability 716 and / or the variant detection classification 718 on the genotype quality metric, and weights each accordingly. The detection integration system 106 further recalibrates the genotype quality metric by increasing or decreasing the value of the genotype quality metric based on the indicated probability.

[0226] As described, the call integration system 106 generates output genotype calls based on the same set of sequencing metrics (or a subset of sequencing metrics shared between the genotype call integration machine learning models 706a and 706b and the call generation model 724). In fact, the call integration system 106 can operate the genotype call integration machine learning models 706a and 706b in parallel with the call generation model 724 to generate metrics for the output genotype calls, genotype probabilities 716, and variant call classifications 718 for recalibrating the generated metrics.

[0227] In one or more specific implementations, the call integration system 106 updates or otherwise modifies the data fields of the merged variant call files 722 according to a specific algorithm. After modifying such data fields, the call integration system 106 can generate a merged variant call file 722 (e.g., a post-filtered variant call file) to include metrics reflecting the updated data fields. For example, in some cases, the call integration system 106 updates the QUAL field for each variant based on the probability of a false positive variant. As indicated above, in some cases, QUAL indicates the probability of a certain variant (or other nucleobase call) at a given position, which is measured in the PHRED scale.

[0228] As mentioned above, in some embodiments, the detection integration system 106 increases or decreases the base call quality metric (e.g., Q score) of the genotype call. Based on the genotype probability 716 and / or the variant call classification 718, for example, the detection integration system 106 increases the base call quality metric of the genotype call that did not pass the quality filter before, and determines that the increased base call quality metric now passes the quality filter. In some such cases, the detection integration system 106 includes genotype calls with such increased base call quality metrics (passing the quality filter) in the post-filtered variant call file. In contrast, in other cases, the detection integration system 106 reduces the base call quality metric of the genotype call that originally passed the quality filter before, and determines that the reduced base call quality metric now fails to pass the quality filter. In some such cases, the call integration system 106 excludes genotype calls with reduced base call quality metrics (failure to pass the quality filter) from the post-filter variant call file, but includes genotype calls with such reduced base call quality metrics in the pre-filter variant call file.

[0229] For example, the detection integration system 106 can remove false positive variant detection and restore false negative variant detection by changing the corresponding base detection quality metric. In order to remove false positives, in some cases, based on the genotype probability 716 and / or variant detection classification 718 from the genotype detection integrated machine learning models 706a and 706b, the detection integration system 106 reduces the base detection quality metric of the genotype detection that initially passes the quality filter. Based on determining that the reduced base detection quality metric drops to below the threshold metric (e.g., Q score 3.0 or 10.0), the detection integration system 106 determines that the genotype detection no longer passes the quality filter. Therefore, the detection integration system 106 filters out or removes the false positive genotype detection by changing the base detection quality metric of the false positive genotype detection that initially passes the filter.

[0230] In addition to removing false positive variant calls based on changes to the base call quality metric, the call integration system 106 can also remove false positive variant calls based on changes to the genotype. In order to remove false positives, in some cases, the call integration system 106 changes the genotype (e.g., GT=1 or 2) of the initial genotype call indicating a nucleobase different from the reference base to the genotype (e.g., GT=0) of the updated genotype call indicating the same nucleobase as the reference base. Based on the genotype being the same as the reference base, the call integration system 106 does not identify the genotype call as a variant, and in some cases excludes the data of the genotype call from the merged variant call file 722. For example, the call integration system 106 may use an empty data indicator for the genotype call (or a specific field) of the merged variant call file 722. In some cases, where a particular sequencing metric does not apply to a particular variant call or VCF field (e.g., where SBS-based calls use a different metric than calls based on assembled nucleotide reads), the call integration system 106 uses a null data indicator.

[0231] When generating the merged variant call file 722, in some embodiments, the call integration system 106 determines the first pipeline accuracy probability of the first pipeline (e.g., based on the first read type) and the second pipeline accuracy probability of the second pipeline (e.g., based on the second read type). To elaborate, the call integration system 106 determines the first pipeline accuracy probability that the first genotype call (e.g., a genotype call generated based on the SBS read) is more accurate than the second genotype call (e.g., a genotype call generated based on the assembled nucleotide read). The call integration system 106 also determines the second pipeline accuracy probability that the second genotype call is more accurate than the first genotype call. In fact, the call integration system 106 can use the genotype call integration machine learning model 706a and / or 706b to determine the possibility or probability that the first genotype call and / or the second genotype call are more accurate. Based on the pipeline accuracy probability, the call integration system 106 can also generate an output genotype call (and the corresponding field in the merged variant call file 722) from the first genotype call and / or the second genotype call.

[0232] To recover the false negative, the call integration system 106 increases the base call quality metric of the genotype call that did not initially pass the quality filter. Based on determining that the increased base call quality metric exceeds the threshold metric, the call integration system 106 determines that the genotype call passes the quality filter. Therefore, the call integration system 106 recovers the false negative genotype call that was initially filtered out by changing the base call quality metric of the false negative genotype call.

[0233] In addition to recovering false negatives based on changes to the base call quality metric, the call integration system 106 can also recover false negative variant calls based on changes to the genotype. In order to recover false negatives, in some cases, the call integration system 106 changes the genotype of the initial genotype call indicating the same nucleobase as the reference base (e.g., GT=0) to a different genotype (e.g., GT=1 or 2) of the updated genotype call indicating a different nucleobase from the reference base. Based on the different genotype of the updated genotype call and the passed base call quality metric, the call integration system 106 identifies the genotype call as a variant and includes the genotype call in the merged variant call file 722.

[0234] In fact, in some specific implementations, the detection integration system 106 uses the detection generation model 724 and the genotype detection integration machine learning models 706a and 706b to operate in a specific sequence order. For example, the detection integration system 106 generates a FASTQ file by converting the BCL file to FASTQ. In addition, the detection integration system 106 (subsequently) uses the mapping and alignment component 710 of the detection generation model 724 to map and align the nucleobases from the sample nucleotide sequence. In some cases, the detection integration system 106 maps and aligns the nucleobases of the sample sequence with respect to the reference sequence 704 (e.g., a reference genome) and / or various alternative support reads.

[0235] After mapping and alignment, as described herein, the call integration system 106 then utilizes the variant caller component 712 of the call generation model 724 to generate initial genotype calls for the sample sequence corresponding to the specific genomic coordinates based on various sequencing metrics. Thereafter or simultaneously, the call integration system 106 also applies the genotype call integration machine learning models 706a and 706b to generate genotype probabilities 716 and variant call classifications 718 based on sequencing metrics extracted via mapping and alignment, variant calls, and / or from other sources as described above. Based on the genotype probabilities 716 and variant call classifications 718, the call integration system 106 recalibrates the genotype calls (e.g., by modifying various data fields corresponding to specific metrics such as QUAL, GT, GQ, GP, and / or PL for nucleobase calls).

[0236] In some cases, the detection integration system 106 further applies a quality filter to the genotype call to determine whether the genotype call passes the quality filter (e.g., a hard pass filter of Q20 or other Q scores). The detection integration system 106 then identifies a subset of genotype calls that represent variants from the reference base and pass the quality filter. The detection integration system 106 further generates a modified or updated variant call file (e.g., a merged variant call file 722), which includes a subset of genotype calls and a recalibrated metric of the subset of genotype calls, such as an updated QUAL metric, an updated GT metric, an updated GQ metric, an updated GP metric, and / or an updated PL metric.

[0237] As mentioned above, in certain described embodiments, the call integration system 106 improves accuracy compared to existing systems. Specifically, the call integration system 106 reduces false positive variant genotype calls and false negative variant genotype calls compared to existing sequencing systems. In fact, by utilizing a genotype call integration machine learning model based on the described sequencing metrics, the call integration system 106 is even improved over previous versions of the call generation model that did not utilize a genotype call integration machine learning model (but still outperforms other systems). Figures 8 to 10B Graphs and tables of experiments demonstrating improved accuracy of the detection integrated system 106 are illustrated.

[0238] For example, Figure 8 The performance of a previous version of a call generation model (e.g., a model that does not utilize a genotype call integration machine learning model) in generating variant calls based on the PrecisionFDA dataset is illustrated. For example, the previous version analyzed the assembled nucleotide reads and SBS reads separately to generate independent results for SNPs and indels. The model generates variant calls for comparison with a benchmark truth value (e.g., from a PrecisionFDA dataset, such as HG001 v4.2.1) to determine performance based on the number of false positives and false negatives.

[0239] As shown, graph 802 corresponds to table 806, and graph 804 corresponds to table 808. Graph 802 depicts a receiver operating characteristic (ROC) curve corresponding to the data of table 806, wherein a previous version of the call generation model (e.g., without machine learning elements) independently determined variant calls for SNPs based on assembled nucleotide reads and SBS reads. Similarly, graph 804 depicts a ROC curve for the data of table 808, wherein the previous version independently determined variant calls for insertions and deletions based on assembled nucleotide reads and SBS reads. Although the performance of the previous system was good in each case (e.g., there were relatively few FPs and FNs compared to other previous systems), the call integration system 106 can still improve the performance by reducing false positives and / or false negatives.

[0240] For example, Fig. 9A A table comparing the performance of a previous version of the call generation model with the performance of the call integration system 106 is illustrated. As shown, table 902 depicts cumulative indications of false positives and false negatives (FP+FN) for a variant call model (SBS+ML+GRAPH), which uses a single read type (e.g., SBS reads) together with machine learning predictions and a graph genome (e.g., Illumina DRAGEN graph reference genome hg19) to generate variant calls for SNPs and indels. Table 902 also depicts results from the call integration system 106, which utilizes a genotype call integration machine learning model to generate variant calls (for SNPs and indels) based on both SBS reads and assembled nucleotide reads (in addition to using specific sequencing metrics and machine learning predictions).

[0241] like Fig. 9A , the experimenter generates the result of table 902 by using different models to generate variant detection for HG002 data set, and the HG002 data set is a specific set of available human genome data for a specific genome sample. In a similar manner, table 904 describes the results of the variant detection of the previous model and the genotype detection integrated machine learning model in generating the HG003 data set. As shown in the figure, the detection integration system 106 with the genotype detection integrated machine learning model is better than the previous model, resulting in less FP+FN metrics in each table, and also has a higher F1 score in each case (for example, for SNP and insertion and deletion in table 902 and table 904). In fact, by utilizing different read sources for different types of reads, compared with a system that cannot handle multiple read types, the detection integration system 106 can generate more accurate variant detection.

[0242] continue Fig. 9B, Table 906 illustrates the results generated by the experimenter when using the genotype call integrated machine learning model to generate SNPs for the HG002 dataset and the HG003 dataset. In addition, Table 908 illustrates the results generated by the experimenter when using the genotype call integrated machine learning model to generate insertions and deletions for the HG002 dataset and the HG003 dataset. In fact, through longer training of the genotype call integrated machine learning model, the experimenter has demonstrated that there is a further improvement in accuracy in addition to the metrics indicated in the previous figures. Compared to existing systems, Fig. 9B The accuracy metrics (e.g., in Table 906 and in Table 908) indicate significant improvements in the accuracy metrics of the genotype calling ensemble machine learning model, particularly in terms of FN, FP, re-call, precision, and F1 measures. In fact, Fig. 9B The accuracy metric for the genotype calling ensemble machine learning model shown in Fig. 9A Improvements in the accuracy metrics of genotype calling ensemble machine learning models in , which is a further improvement over existing systems that do not use genotype calling ensemble machine learning models.

[0243] like Fig. 10A As illustrated in , Figure 1002 depicts a ROC curve for comparing the performance of different variant detectors in generating variant detection for SNP. For example, Figure 1002 shows that the curve with the largest area under the curve usually performs the best ROC curve. As shown in the figure, the detection integration system 106 with the genotype detection integrated machine learning model is superior to other models. Other models include SBS+ML+GRAPH model (as reflected in the table of Figure 9), (only) a model for generating variant detection from assembled nucleotide reads (for example, without further analysis or machine learning techniques) and (only) a model for generating variant detection from SBS reads (for example, without further analysis or machine learning techniques). As indicated in Figure 1002, the genotype detection integrated machine learning model has the highest area under the curve and the least false positives, and is superior to other models on a test data set (such as a PrecisionFDA data set).

[0244] Fig. 10B 1002 consistent with the column diagram 1004 is illustrated. In detail, the column diagram 1004 provides a genotype detection integrated machine learning model and the SBS+ML+GRAPH model for SNP (e.g., for chromosome 20-21-22) in the variant detection of the alternative visualization of comparison. In fact, the column diagram 1004 indicates false negatives and false positives and their cumulative totals for each model. As shown in the figure, the genotype detection integrated machine learning model generates more accurate variant detection than the SBS+ML+GRAPH model, resulting in less false negatives, less false positives and less overall FP+FN.

[0245] Now turn to Fig.11 , which illustrates an example flow chart of a series of actions for generating output genotype calls using a genotype call integration machine learning model according to one or more embodiments. Fig.11 The actions according to one embodiment are illustrated, but alternative embodiments may omit, add, reorder, and / or modify Fig.11 Any action shown. Fig.11 The actions of may be performed as part of a method. Alternatively, the non-transitory computer-readable storage medium may include a computer program that, when executed by one or more processors, causes the computing device to perform Fig.11 In some further embodiments, the system includes at least one processor and a non-transitory computer-readable medium, which includes instructions that, when executed by the one or more processors, cause the system to perform Fig.11 instructions for actions.

[0246] Fig.11 A series of actions 1100 for generating output genotype calls using a genotype call integrated machine learning model are illustrated. Specifically, the series of actions 1100 include receiving a first genotype call for a first read type and an action 1102 for a second genotype call for a second read type. For example, action 1102 may involve receiving a first genotype call for a first type of nucleotide read corresponding to a first threshold number of nucleobases and a second genotype call for a second type of nucleotide read corresponding to a second threshold number of nucleobases for one or more genomic coordinates of a genomic sample. The first type of nucleotide read may include a nucleotide read synthesized from a sample library fragment shorter than the first threshold number of nucleobases. The second type of nucleotide read may include an assembled nucleotide read that has been assembled by a shorter nucleotide read to form a continuous sequence that satisfies a first threshold number of nucleobases, a cycle common sequencing (CCS) read that satisfies a first threshold number of nucleobases, or a nanopore long read that satisfies a first threshold number of nucleobases. The first genotype call may include a first variant call or a first reference call. The second genotype call may include a second variant call or a second reference call. In some cases, the first genotype call or the second genotype call includes a null data indicator.

[0247] like Fig.11As further illustrated in , the series of actions 1100 includes an action 1104 of identifying a sequencing metric. Specifically, action 1104 may involve identifying a sequencing metric corresponding to a first genotype call or a second genotype call. For example, action 1104 involves identifying a sequencing metric corresponding to a first genotype call or a second genotype call by identifying one or more of the following, a first set of sequencing metrics associated with a first genotype call corresponding to a nucleotide read of a first type, a second set of sequencing metrics associated with a second genotype call corresponding to a nucleotide read of a second type, or a shared set of sequencing metrics associated with a first genotype call and a second genotype call. In some cases, action 1104 involves identifying a sequencing metric corresponding to a first genotype call or a second genotype call by determining one or more of the following: a read-based sequencing metric, a call model-generated sequencing metric, a sequencing metric from an external source, or a second read type sequencing metric associated with a second genotype call corresponding to a nucleotide read of a second type.

[0248] In one or more embodiments, action 1104 involves identifying a read-based sequencing metric that includes one or more of the following: an allele frequency corresponding to an allele called by a first genotype, an allele called by a second genotype, or a different allele called by an alternative genotype that is different from the first genotype call and the second genotype call; a depth of coverage of a first type of nucleotide read corresponding to the first genotype call or a second type of nucleotide read corresponding to the second genotype call; an average depth of coverage of a first type of nucleotide read corresponding to the first genotype call or a second type of nucleotide read corresponding to the second genotype call; a mapping quality metric of a first type of nucleotide read corresponding to the first genotype call or a second type of nucleotide read corresponding to the second genotype call; or a nucleobase composition of one or more nucleotide reads from the first type of nucleotide read or the second type of nucleotide read.

[0249] In certain embodiments, act 1104 involves identifying sequencing metrics generated by the calling model, which include one or more of the following: a genotype metric for a first genotype call determined based on a first type of nucleotide read or a second genotype call determined based on a second type of nucleotide read, a base call quality metric, a genotype quality metric, a genotype probability metric, a genotype likelihood metric (e.g., a non-PHRED scaled likelihood metric or a PHRED scaled likelihood metric).

[0250] In these or other embodiments, act 1104 involves identifying sequencing metrics from an external source that include one or more of the following: a mappability metric that indicates the difficulty of mapping a nucleotide read to one or more genomic coordinates within a reference genome; a guanine-cytosine content metric that indicates counts of guanine-cytosine content corresponding to one or more genomic coordinates within the reference genome; a confidence classification or confidence score that indicates the degree to which the nucleobase at the one or more genomic coordinates can be accurately determined; a repeat classification that indicates the category of a repetitive genomic region for the one or more genomic coordinates; an indicator that the one or more genomic coordinates are part of a cytosine quadruplex (C-quadruplex) within the reference genome; an indicator that the one or more genomic coordinates are part of a guanine quadruplex (G-quadruplex) within the reference genome; or an indicator that the one or more genomic coordinates are part of a homopolymer within the reference genome.

[0251] Additionally, the series of actions 1100 may include an action 1106 of generating genotype probabilities and / or variant call classifications using a genotype call integrated machine learning model. Specifically, action 1106 may involve generating genotype probabilities for genotype calls for one or more genomic coordinates using a genotype call integrated machine learning model and based on sequencing metrics. In some cases, action 1106 involves generating variant call classifications for candidate variant calls at one or more genomic coordinates using a genotype call integrated machine learning model and based on sequencing metrics.

[0252] In one or more embodiments, action 1106 involves generating genotype probabilities by generating genotype probabilities of one or more candidate SNPs using a genotype call ensemble machine learning model trained with single nucleotide polymorphism (SNP) training data. In certain embodiments, action 1106 involves: generating a first genotype probability for a genomic sample including a homozygous reference genotype at one or more genomic coordinates; generating a second genotype probability for a genomic sample including a heterozygous variant genotype at one or more genomic coordinates; and generating a third genotype probability for a genomic sample including a homozygous variant genotype at one or more genomic coordinates.

[0253] In certain embodiments, action 1106 involves generating a variant call classification for one or more candidate insertions or deletions (indels) using a genotype call ensemble machine learning model trained with indel training data. Action 1106 may involve generating a variant call classification for a candidate variant by generating one or more of: a first true positive variant probability that a first genotype call constitutes a true positive variant for one or more genomic coordinates; a second true positive variant probability that a second genotype call constitutes a true positive variant for one or more genomic coordinates; a first zygosity error probability that a first genotype call includes a genotypic zygosity error at one or more genomic coordinates; a second genotype call includes a second zygosity error probability for a genotypic zygosity error at one or more genomic coordinates; or a reference probability of a homozygous reference genotype at one or more genomic coordinates.

[0254] In some embodiments, the series of actions 1108 includes an action 1108 of generating an output genotype call from a genotype probability and / or a variant call classification. Specifically, action 1108 may involve generating an output genotype call for one or more genomic coordinates of a genomic sample based on the genotype probability. In some cases, action 1108 involves generating an output genotype call for one or more genomic coordinates of a genomic sample based on a variant call classification. In certain embodiments, action 1108 involves generating an output genotype call indicating the presence or absence of a SNP at one or more genomic coordinates of a genomic sample. In some embodiments, action 1108 involves generating an output genotype call indicating the presence or absence of an insertion or deletion at one or more genomic coordinates of a genomic sample. Action 1108 may include: selecting a first genotype call or a second genotype call; or generating a different genotype call that is different from the first genotype call and the second genotype call.

[0255] In certain embodiments, action 1108 involves selecting a first genotype call instead of a second genotype call. Selecting a first genotype call instead of a second genotype call can involve: selecting a homozygous reference genotype call from the first genotype call instead of a heterozygous variant genotype call or a homozygous variant genotype call from the second genotype call; selecting a heterozygous variant genotype call from the second genotype call instead of a homozygous reference genotype call or a homozygous variant genotype call from the first genotype call; or selecting a homozygous variant genotype call from the first genotype call instead of a heterozygous variant genotype call or a homozygous reference genotype call from the second genotype call.

[0256] In some cases, action 1108 involves selecting a second genotype call instead of the first genotype call by: selecting a homozygous reference genotype call from the second genotype call instead of selecting a heterozygous variant genotype call or a homozygous variant genotype call from the first genotype call; selecting a heterozygous variant genotype call from the second genotype call instead of selecting a homozygous reference genotype call or a homozygous variant genotype call from the first genotype call; or selecting a homozygous variant genotype call from the second genotype call instead of selecting a heterozygous variant genotype call or a homozygous reference genotype call from the first genotype call. Action 1108 may involve: selecting the first genotype call or the second genotype call; or generating a different genotype call that is different from the first genotype call and the second genotype call.

[0257] In one or more embodiments, the series of acts 1100 includes an act of modifying a genotype metric, a base call quality metric, a genotype quality metric, a genotype probability metric, a genotype likelihood metric, or a PHRED scaled genotype likelihood metric based on the genotype probability and / or the variant call classification. In these or other embodiments, the series of acts 1100 includes an act of generating a variant call file that includes the modified genotype metric, the modified base call quality metric, the modified genotype quality metric, the modified genotype probability metric, the modified genotype likelihood metric, or the modified PHRED scaled genotype likelihood metric.

[0258] In certain embodiments, the series of actions 1100 includes an action of receiving a first genotype call by receiving the first genotype call based on the first type of nucleotide reads as part of a first variant call file. In the same or other embodiments, the series of actions 1100 includes the following actions: receiving a second genotype call by receiving the second genotype call based on the second type of nucleotide reads as part of a second variant call file; and generating a merged variant call file including the first genotype call or the second genotype call.

[0259] In some embodiments, the series of actions 1100 includes an action of determining that the first genotype call includes a first alternative nucleobase that is different from the second alternative nucleobase of the second genotype call. The series of actions 1100 may also include an action of integrating a machine learning model with the genotype calls and generating a first pipeline accuracy probability that the first genotype call is more accurate than the second genotype call and a second pipeline accuracy probability that the second genotype call is more accurate than the first genotype call based on a sequencing metric. In addition, the series of actions 1100 may include an action of generating an output genotype call by selecting a first genotype call or a second genotype call for one or more genomic coordinates of a genomic sample based on the first pipeline accuracy probability and the second pipeline accuracy probability.

[0260] The series of actions 1100 may include an action of determining that the first true positive variant probability fails to meet the likelihood threshold. Additionally, the series of actions 1100 may include an action of generating or utilizing a second true positive variant probability 1100 based on determining that the first true positive variant probability fails to meet the likelihood threshold.

[0261] Methods described herein can be used in combination with multiple nucleic acid sequencing techniques. Particularly suitable techniques are those in which nucleic acids are attached to fixed positions in an array so that their relative positions do not change and in which the array is repeatedly imaged. Embodiments in which images are obtained in different color channels (e.g., matching different markers for distinguishing a type of core base from another type of core base) are particularly suitable. In some embodiments, the process for determining the nucleotide sequence of a target nucleic acid (i.e., nucleic acid polymer) can be an automated process. Preferred embodiments include sequencing by synthesis (SBS) technology.

[0262] SBS techniques generally include enzymatic extension of nascent nucleic acid chains by repeated addition of nucleotides to the template strand. In traditional SBS methods, a single nucleotide monomer can be provided to the target nucleotide in the presence of a polymerase in each delivery. However, in the methods described herein, more than one type of nucleotide monomer can be provided to the target nucleic acid in the presence of a polymerase in the delivery.

[0263] SBS can utilize nucleotide monomers with terminator parts or lack nucleotide monomers of any terminator parts.The method utilizing the nucleotide monomers lacking terminator includes for example pyrophosphate sequencing and the sequencing of nucleotides using gamma-phosphate labeling, as described in further detail below.In the method using the nucleotide monomers lacking terminator, the number of nucleotides added in each cycle is usually variable, and the number depends on the mode of template sequence and nucleotide delivery.For utilizing the SBS technology of the nucleotide monomers with terminator parts, terminator can be effectively irreversible under the sequencing conditions used, such as the situation of traditional Sanger sequencing utilizing dideoxynucleotides, or terminator can be reversible, such as the situation of the sequencing method developed by Solexa (now Illumina, Inc.).

[0264] The SBS technique can utilize nucleotide monomers with a labeling portion or nucleotide monomers lacking a labeling portion. Thus, incorporation events can be detected based on the following items: properties of the label, such as the fluorescence of the label; properties of the nucleotide monomer, such as molecular weight or charge; byproducts of the incorporated nucleotide, such as the release of pyrophosphate; and the like. In embodiments where two or more different nucleotides are present in the sequencing reagent, the different nucleotides may be distinguishable from each other, or alternatively, the two or more different labels may be indistinguishable under the detection technology used. For example, different nucleotides present in the sequencing reagent may have different labels, and they may be distinguished using appropriate optical devices, as exemplified by the sequencing method developed by Solexa (now Illumina, Inc.).

[0265] Preferred embodiments include pyrophosphate sequencing technology. Pyrophosphate sequencing detects the release of inorganic pyrophosphate (Ppi) when a specific nucleotide is incorporated into a nascent chain (Ronaghi, M., Karamohamed, S., Pettersson, B., Uhlen, M. and Nyren, P. (1996) "Real-time DNA sequencing using detection of pyrophosphate release." Analytical Biochemistry 242 (1), 84-9; Ronaghi, M. (2001) "Pyrosequencing sheds light on DNA sequencing" Genome Res. 11 (1), 3-11; Ronaghi, M., Uhlen, M. and Nyren, P. (1998), "A sequencing method based on real-time pyrophosphate." Science 281 (5375), 363; U.S. Pat. No. 6,210,891; U.S. Pat. No. 6,258,568 and U.S. Pat. No. 6,274,320, the disclosures of which are incorporated herein by reference in their entirety). In pyrophosphate sequencing, the released Ppi can be immediately converted into ATP by adenosine triphosphate (ATP) sulfurylase for detection, and the generated ATP level is detected via photons generated by luciferase. The nucleic acid to be sequenced can be attached to a feature in an array, and the array can be imaged to capture the chemiluminescent signal generated by the incorporation of nucleotides at the feature of the array. The image can be obtained after the array is treated with a specific nucleotide type (e.g., A, T, C or G). The images obtained after adding each nucleotide type will differ in which features are detected in the array. These differences in the image reflect the different sequence contents of the features on the array. However, the relative position of each feature will remain unchanged in the image. The images can be stored, processed and analyzed using the methods described herein. For example, images obtained after treating the array with each different nucleotide type can be processed in the same manner as exemplified herein for images obtained from different detection channels for a reversible terminator-based sequencing method.

[0266] In another exemplary type of SBS, cycle sequencing is accomplished by the stepwise addition of reversible terminator nucleotides, which contain, for example, cleavable or photobleachable dye labels, as described in, for example, WO 04 / 018497 and U.S. Patent No. 7,057,026, the disclosures of which are incorporated herein by reference. The method is commercialized by Solexa (now Illumina Inc.) and is also described in WO 91 / 06678 and WO 07 / 123,744, the disclosures of each of which are incorporated herein by reference. The availability of fluorescently labeled terminators, in which termination can be reversible and the fluorescent label can be cleaved, facilitates efficient cyclic reversible termination (CRT) sequencing. Polymerases can also be co-engineered to efficiently incorporate these modified nucleotides and extend from these modified nucleotides.

[0267] Preferably, in the sequencing embodiment based on reversible terminator, the label does not substantially inhibit extension under SBS reaction conditions. However, the detection label can be removable, such as removed by cleavage or degradation. Images can be captured after the label is incorporated into the arrayed nucleic acid feature. In a specific embodiment, each cycle involves four different nucleotide types being delivered to the array simultaneously, and each nucleotide type has a label that is spectrally different. Four images can then be obtained, each image using a detection channel that is selective to one label in the four different labels. Alternatively, different nucleotide types can be added sequentially, and the image of the array can be obtained between each addition step. In such embodiments, each image will show the nucleic acid feature that has been incorporated with a specific type of nucleotide. Due to the different sequence content of each feature, different feature portions are present or absent in different images. However, the relative position of the feature portion will remain unchanged in the image. The image obtained by such reversible terminator-SBS method can be stored, processed and analyzed as described herein. After the image capture step, the label can be removed and the reversible terminator portion can be removed for subsequent cycles of nucleotide addition and detection. Removing labels after they have been detected in a particular cycle and before subsequent cycles can provide the advantage of reducing background signal and crosstalk between cycles. Examples of useful labeling and removal methods are set forth below.

[0268] In a specific embodiment, some or all of the nucleotide monomers may include a reversible terminator. In such embodiments, the reversible terminator / cleavable fluorophore may include a fluorophore (Metzker, Genome Res. 15: 1767-1776 (2005), which is incorporated herein by reference) connected to a ribose moiety via a 3' ester bond. Other methods have separated terminator chemistry from the cleavage of a fluorescent label (Ruparel et al., Proc Natl Acad Sci USA 102: 5932-7 (2005), which is incorporated herein by reference in its entirety). Ruparel et al. describe the development of reversible terminators that use small 3' allyl groups to block extension, but can be easily deblocked by a short treatment with a palladium catalyst. The fluorophore is attached to the base via a photocleavable linker that can be easily cleaved by exposure to long wavelength ultraviolet light for 30 seconds. Therefore, disulfide reduction or photocleavage can be used as a cleavable linker. Another method of reversible termination is to use natural termination, which occurs after a bulky dye is placed on a dNTP. The presence of a charged bulky dye on a dNTP can act as an efficient terminator by steric hindrance and / or electrostatic hindrance. Unless the dye is removed, the presence of an incorporation event prevents further incorporation. The cracking of the dye removes the fluorophore and effectively reverses termination. The example of modified nucleotides is also described in U.S. Patent No. 7,427,673 and U.S. Patent No. 7,057,026, the disclosure of which is incorporated herein by reference in its entirety.

[0269] Additional exemplary SBS systems and methods that may be utilized with the methods and systems described herein are described in U.S. Patent Application Publication No. 2007 / 0166705, U.S. Patent Application Publication No. 2006 / 0188901, U.S. Patent No. 7,057,026, U.S. Patent Application Publication No. 2006 / 0240439, U.S. Patent Application Publication No. 2006 / 0281109, PCT Publication No. WO05 / 065814, U.S. Patent Application Publication No. 2005 / 0100900, PCT Publication No. WO 06 / 064199, PCT Publication No. WO07 / 010,251, U.S. Patent Application Publication No. 2012 / 0270305, and U.S. Patent Application Publication No. 2013 / 0260372, the disclosures of which are incorporated herein by reference in their entireties.

[0270] Some embodiments can use the detection of four different nucleotides using less than four different marks.For example, the method and system described in the material of the U.S. Patent Application Publication No. 2013 / 0079232 incorporated can be used to perform SBS.As a first example, a pair of nucleotide types can be detected at the same wavelength, but based on the intensity difference of a member in the pair relative to another member, or based on a member in the pair causing a change (for example, by chemical modification, photochemical modification or physical modification) that a signal that is obvious compared with the signal of another member of the pair detected appears or disappears.As a second example, three of the four different nucleotide types can be detected under specific conditions, and the fourth nucleotide type lacks a mark (for example, the minimum detection caused by background fluorescence, etc.) that can be detected under those conditions or detected minimally under those conditions.The first three nucleotide types can be determined to be incorporated into nucleic acid based on the existence of its corresponding signal, and the fourth nucleotide type can be determined to be incorporated into nucleic acid based on the absence of any signal or the minimum detection of any signal.As a third example, a nucleotide type can include a mark detected in two different channels, and other nucleotide types are detected in no more than one channel. The three exemplary configurations described above are not considered mutually exclusive and can be used in various combinations. An exemplary embodiment combining all three examples is a fluorescence-based SBS method that uses a first nucleotide type detected in a first channel (e.g., dATP with a label detected in the first channel when excited by a first excitation wavelength), a second nucleotide type detected in a second channel (e.g., dCTP with a label detected in the second channel when excited by a second excitation wavelength), a third nucleotide type detected in both the first channel and the second channel (e.g., dTTP with at least one label detected in both channels when excited by the first excitation wavelength and / or the second excitation wavelength), and a fourth nucleotide type that lacks a label detected or minimally detected in either channel (e.g., dGTP without a label).

[0271] In addition, as described in the materials of incorporated U.S. Patent Application Publication No. 2013 / 0079232, a single channel can be used to obtain sequencing data. In such so-called single dye sequencing methods, a first nucleotide type is marked, but the mark is removed after the first image is generated, and a second nucleotide type is marked only after the first image is generated. A third nucleotide type retains its mark in both the first image and the second image, and a fourth nucleotide type remains unmarked in both images.

[0272] Some embodiments can utilize sequencing technology while connecting. Such technology utilizes DNA ligase to incorporate oligonucleotides and recognize the incorporation of such oligonucleotides. Oligonucleotides generally have different labels related to the identity of specific nucleotides in the sequence of oligonucleotide hybridization. As with other SBS methods, images can be obtained after the array of nucleic acid features is treated with labeled sequencing reagents. Each image will show the nucleic acid features that have been incorporated with a specific type of label. Due to the different sequence content of each feature, different features are present or absent in different images, but the relative position of the features will remain unchanged in the image. The images obtained by the sequencing method based on connection can be stored, processed and analyzed as described herein. Exemplary SBS systems and methods that can be used with the methods and systems described herein are described in U.S. Patent No. 6,969,488, U.S. Patent No. 6,172,218 and U.S. Patent No. 6,306,597, and the disclosures of these patents are incorporated herein by reference in their entirety.

[0273] Some embodiments may utilize nanopore sequencing (Deamer, DW and Akeson, M. "Nanopores and nucleic acids: prospects for ultrarapid sequencing." Trends Biotechnol. 18, 147-151 (2000); Deamer, D. and D. Branton, "Characterization of nucleic acids by nanopore analysis". Acc. Chem. Res. 35: 817-825 (2002); Li, J., M. Gershow, D. Stein, E. Brandin and JA Golovchenko, "DNA molecules and configurations in a solid-state nanopore microscope" Nat. Mater. 2: 611-615 (2003), the disclosures of which are incorporated herein by reference in their entirety). In such embodiments, the target nucleic acid passes through a nanopore. The nanopore may be a synthetic pore or a biological membrane protein, such as α-hemolysin. As the target nucleic acid passes through the nanopore, each base pair can be identified by measuring the fluctuations in the conductivity of the pore. (U.S. Pat. No. 7,001,792; Soni, GV and Meller, "A. Progress toward ultrafast DNA sequencing using solid-state nanopores." Clin. Chem. 53, 1996-2001 (2007); Healy, K. "Nanopore-based single-molecule DNA analysis." Nanomed., 2, 459-481 (2007); Cockroft, SL, Chu, J., Amorin, M. and Ghadiri, MR "A single-molecule nanopore device detects DNA polymerase activity with single-nucleotide resolution." J. Am. Chem. Soc. 130, 818-820 (2008), the disclosures of which are incorporated herein by reference in their entirety). The data obtained from nanopore sequencing can be stored, processed and analyzed as described herein. In particular, according to the exemplary processing of optical images and other images described herein, the data can be processed like an image.

[0274] Some embodiments may utilize methods involving real-time monitoring of DNA polymerase activity. Nucleotide incorporation may be detected by fluorescence resonance energy transfer (FRET) interactions between a polymerase carrying a fluorophore and a γ-phosphate labeled nucleotide, as described, for example, in U.S. Pat. No. 7,329,492 and U.S. Pat. No. 7,211,414 (each of which is incorporated herein by reference), or may be detected by zero-mode waveguides, as described, for example, in U.S. Pat. No. 7,315,019 (which is incorporated herein by reference), and may be detected by fluorescent nucleotide analogs and engineered polymerases, as described, for example, in U.S. Pat. No. 7,405,281 and U.S. Patent Application Publication No. 2008 / 0108082 (each of which is incorporated herein by reference). Illumination can be limited to a volume of the order of magnitude surrounding the surface-tethered polymerase, so that the incorporation of fluorescently labeled nucleotides can be observed under low background (Levene, MJ et al., "Zero-mode waveguides for single-molecule analysis at high concentrations." Science 299, 682-686 (2003); Lundquist, PM et al., "Parallel confocal detection of single molecules in real time." Opt. Lett. 33, 1026-1028 (2008); Korlach, J. et al., "Selective aluminum passivation for targeted immobilization of single DNA polymerase molecules in zero-mode waveguide nano structures." Proc. Natl. Acad. Sci. USA 105, 1176-1181 (2008), the disclosures of which are incorporated herein by reference in their entirety). Images obtained by such methods can be stored, processed, and analyzed as described herein.

[0275] Some SBS embodiments include detecting the proton released when nucleotides are incorporated into extension products. For example, sequencing based on the detection of releasing protons can use the commercially available electrical detectors and related technologies available from Ion Torrent company (Guilford, CT, it is a Life Technologies subsidiary) or the sequencing methods and systems described in US2009 / 0026082A1, US2009 / 0127589A1, US2010 / 0137143A1 or US 2010 / 0282617A1, each of which is incorporated herein by reference. The method for amplifying target nucleic acids using kinetic exclusion set forth herein can be easily applied to substrates for detecting protons. More specifically, the method set forth herein can be used to produce amplicon clone colonies for detecting protons.

[0276] The above-mentioned SBS method can be advantageously carried out in a variety of formats so that a plurality of different target nucleic acids are manipulated simultaneously. In a specific embodiment, different target nucleic acids can be processed in a common reaction vessel or on the surface of a specific substrate. This allows to conveniently deliver sequencing reagents, remove unreacted reagents and detect incorporation events in a variety of ways. In the embodiment using the target nucleic acid of surface binding, the target nucleic acid can be an array format. In the array format, the target nucleic acid can be attached to the surface in a spatially distinguishable manner generally. The target nucleic acid can be attached by direct covalent attachment, attached to beads or other particles or attached to a polymerase or other molecules attached to the surface to combine. The array can include a single copy of the target nucleic acid at each site (also referred to as a feature portion), or multiple copies with the same sequence can be present at each site or feature portion. Multiple copies can be produced by an amplification method (such as bridge amplification or emulsion PCR as described in further detail below).

[0277] The methods described herein may use arrays having features at any of a variety of densities, including, for example, at least about 10 features / cm2, 100 features / cm2, 500 features / cm2, 1,000 features / cm2, 5,000 features / cm2, 10,000 features / cm2, 50,000 features / cm2, 100,000 features / cm2, 1,000,000 features / cm2, 5,000,000 features / cm2, or more.

[0278] The advantage of the method set forth herein is that they provide rapid and effective detection of multiple target nucleic acids in parallel. Therefore, the present disclosure provides an integrated system that can use technology known in the art (such as those illustrated above) to prepare and detect nucleic acids. Therefore, the integrated system of the present disclosure may include a fluid assembly that can deliver amplification reagents and / or sequencing reagents to one or more fixed DNA fragments, and the system includes components such as pumps, valves, reservoirs, fluid pipelines, etc. The circulation cell can be configured for and / or for detecting target nucleic acids in the integrated system. Exemplary circulation cells are described in, for example, US2010 / 0111768A1 and U.S. Serial No. 13 / 273,666, each of which is incorporated herein by reference. As illustrated for the circulation cell, one or more fluid assemblies of the integrated system can be used for amplification method and detection method. Taking nucleic acid sequencing embodiment as an example, one or more fluid assemblies of the integrated system can be used for the amplification method set forth herein and for delivering sequencing reagents in sequencing method (such as those illustrated above). Alternatively, the integrated system may include a separate fluid system to perform amplification method and perform detection method. Examples of integrated sequencing systems capable of producing amplified nucleic acids and also determining nucleic acid sequences include, but are not limited to, MiSeq TM The ELISA platform (Illumina, Inc., San Diego, CA) and the apparatus described in U.S. Ser. No. 13 / 273,666, which is incorporated herein by reference.

[0279] The above-mentioned sequencing system sequences the nucleic acid polymers present in the sample received by the sequencing device. As defined herein, "sample" and its derivatives are used in their broadest sense, including any specimens, cultures, etc. suspected of containing a target. In some embodiments, the sample includes nucleic acids in DNA, RNA, PNA, LNA, chimeric or hybrid forms. The sample may include any biological, clinical, surgical, agricultural, atmospheric or aquatic animal and plant specimens containing one or more nucleic acids. The term also includes any isolated nucleic acid sample, such as genomic DNA, fresh frozen or formalin-fixed paraffin-embedded nucleic acid specimens. It is also envisioned that the source of the sample can be: a single individual, a collection of nucleic acid samples from genetically related members, a nucleic acid sample from a genetically unrelated member, a nucleic acid sample (matched therewith) from a single individual (such as a tumor sample and a normal tissue sample), or a sample from a single source containing two different forms of genetic material (such as maternal DNA and fetal DNA obtained from a maternal subject), or there is contaminating bacterial DNA in a sample containing plant or animal DNA. In some embodiments, the source of the nucleic acid material may include nucleic acids obtained from a newborn, such as nucleic acids commonly used for newborn screening.

[0280] Nucleic acid samples may include high molecular weight substances, such as genomic DNA (gDNA). Samples may include low molecular weight substances, such as nucleic acid molecules obtained from FFPE samples or archived DNA samples. In another embodiment, low molecular weight substances include DNA of enzymatic fragmentation or mechanical fragmentation. Samples may include cell-free circulating DNA. In some embodiments, samples may include nucleic acid molecules obtained from biopsy tissue, tumor, scrapings, swabs, blood, mucus, urine, blood plasma, semen, hair, laser capture microdissection, surgical resection and other clinical or laboratory samples. In some embodiments, samples may be epidemiological samples, agricultural samples, forensic samples or pathogenic samples. In some embodiments, samples may include nucleic acid molecules obtained from animals (such as humans or mammalian sources). In another embodiment, samples may include nucleic acid molecules obtained from non-mammalian sources (such as plants, bacteria, viruses or fungi). In some embodiments, the source of nucleic acid molecules may be archived or extinct samples or species.

[0281] In addition, the methods and compositions disclosed herein can be used to amplify nucleic acid samples with low-quality nucleic acid molecules, such as degraded and / or fragmented genomic DNA from forensic samples. In one embodiment, forensic samples may include nucleic acids obtained from crime scenes, nucleic acids obtained from missing persons DNA databases, nucleic acids obtained from laboratories associated with forensic investigations, or include forensic samples obtained by law enforcement agencies, one or more military services, or any such personnel. Nucleic acid samples can be purified samples or lysates containing crude DNA, such as from oral swabs, paper, fabrics, or other substrates that can be impregnated with saliva, blood, or other body fluids. Therefore, in some embodiments, the nucleic acid sample may include a small amount of DNA (such as genomic DNA), or a fragmented portion of DNA. In some embodiments, the target sequence may be present in one or more body fluids, wherein body fluids include but are not limited to blood, sputum, plasma, semen, urine, and serum. In some embodiments, the target sequence may be obtained from hair, skin, tissue samples, autopsy, or remains of victims. In some embodiments, nucleic acids comprising one or more target sequences may be obtained from dead animals or humans. In some embodiments, the target sequence may include a nucleic acid obtained from non-human DNA (such as microbial, plant, or insect DNA). In some embodiments, the target sequence or amplified target sequence is directed to the purpose of human identification. In some embodiments, the disclosure as a whole relates to methods for identifying the characteristics of forensic samples. In some embodiments, the disclosure as a whole relates to human identification methods using one or more target-specific primers disclosed herein or one or more target-specific primers designed using the primer design standards outlined herein. In one embodiment, a forensic sample or human identification sample containing at least one target sequence can be amplified using any one or more target-specific primers disclosed herein or using the primer standards outlined herein.

[0282] The components of the detection integration system 106 may include software, hardware, or both. For example, the components of the detection integration system 106 may include one or more instructions that are stored on a computer-readable storage medium and can be executed by a processor of one or more computing devices (e.g., client device 108). When executed by one or more processors, the computer-executable instructions of the detection integration system 106 may cause the computing device to perform the bubble detection method described herein. Alternatively, the components of the detection integration system 106 may include hardware, such as a dedicated processing device for performing a certain function or group of functions. Additionally or alternatively, the components of the detection integration system 106 may include a combination of computer-executable instructions and hardware.

[0283] In addition, the components of the detection integration system 106 that perform the functions described herein with respect to the detection integration system 106 may be implemented, for example, as part of a stand-alone application, a module of an application, a plug-in of an application, one or more library functions that can be detected by other applications, and / or a cloud computing model. Therefore, the components of the detection integration system 106 may be implemented as part of a stand-alone application on a personal computing device or a mobile device. Additionally or alternatively, the components of the detection integration system 106 may be implemented in any application that provides sequencing services, including but not limited to Illumina BaseSpace, Illumina DRAGEN, or Illumina TruSight software. "Illumina", "BaseSpace", "DRAGEN", and "TruSight" are registered trademarks or trademarks of Illumina, Inc. in the United States and / or other countries.

[0284] As discussed in more detail below, embodiments of the present disclosure may include or utilize a special-purpose or general-purpose computer including computer hardware (such as, for example, one or more processors and system memory). Embodiments within the scope of the present disclosure also include physical and other computer-readable media for carrying or storing computer-executable instructions and / or data structures. Specifically, one or more of the processes described herein may be implemented at least in part as instructions embodied in a non-transitory computer-readable medium and executable by one or more computing devices (e.g., any of the media content access devices described herein). In general, a processor (e.g., a microprocessor) receives instructions from a non-transitory computer-readable medium (e.g., a memory, etc.) and executes those instructions, thereby performing one or more processes, including one or more of the processes described herein.

[0285] Computer readable media can be any available media that can be accessed by a general or special purpose computer system. A computer readable medium that stores computer executable instructions is a non-transitory computer readable storage medium (device). A computer readable medium that carries computer executable instructions is a transmission medium. Therefore, by way of example and not limitation, embodiments of the present disclosure may include at least two distinct types of computer readable media: a non-transitory computer readable storage medium (device) and a transmission medium.

[0286] Non-transitory computer-readable storage media (devices) include RAM, ROM, EEPROM, CD-ROM, solid-state drives (SSD) (e.g., RAM-based), flash memory, phase-change memory (PCM), other types of memory, other optical disk storage devices, magnetic disk storage devices or other magnetic storage devices, or any other media that can be used to store desired program code means in the form of computer-executable instructions or data structures and which can be accessed by a general or special purpose computer.

[0287] "Network" is defined as one or more data links that enable the transmission of electronic data between computer systems and / or modules and / or other electronic devices. When information is transferred or provided to a computer via a network or another communication connection (hardwired, wireless, or a combination of hardwired or wireless), the computer appropriately regards the connection as a transmission medium. The transmission medium may include a network and / or data link, which can be used to carry the desired program code means in the form of computer executable instructions or data structures, and which can be accessed by general or special computers. The above combinations should also be included in the scope of computer-readable media.

[0288] Furthermore, upon reaching various computer system components, program code means in the form of computer executable instructions or data structures may be automatically transferred from a transmission medium to a non-transitory computer readable storage medium (device) (or vice versa). For example, a computer executable instruction or data structure received over a network or data link may be buffered in RAM within a network interface module (e.g., NIC), and then ultimately transferred to the computer system RAM and / or to a less volatile computer storage medium (device) at the computer system. Thus, it should be understood that a non-transitory computer readable storage medium (device) may be included in a computer system component that also (or even primarily) utilizes a transmission medium.

[0289] Computer executable instructions include, for example, instructions and data that make a general-purpose computer, a special-purpose computer, or a special-purpose processing device perform certain functions or groups of functions when executed at a processor. In some embodiments, computer executable instructions are executed on a general-purpose computer to turn a general-purpose computer into a special-purpose computer that implements an element of the present disclosure. Computer executable instructions can be, for example, binary numbers, intermediate format instructions such as assembly language, or even source code. Although the subject matter has been described in a language specific to structural features and / or method actions, it should be understood that the subject matter defined in the appended claims is not necessarily limited to the described features or actions. On the contrary, the described features and actions are disclosed as exemplary forms of implementing the claims.

[0290] Those skilled in the art will appreciate that the present disclosure can be practiced in a network computing environment with many types of computer system configurations, including personal computers, desktop computers, portable computers, message processors, handheld devices, multiprocessor systems, microprocessor-based or programmable consumer electronics, network PCs, minicomputers, mainframe computers, mobile phones, PDAs, tablet computers, pagers, routers, switches, etc. The present disclosure can also be practiced in a distributed system environment, where both local and remote computer systems linked by a network (by a hardwired data link, a wireless data link, or a combination of hardwired and wireless data links) perform tasks. In a distributed system environment, program modules can be located in both local and remote memory storage devices.

[0291] Embodiments of the present disclosure may also be implemented in a cloud computing environment. In this specification, "cloud computing" is defined as a model for implementing on-demand network access to a shared pool of configurable computing resources. For example, cloud computing may be employed in the marketplace to provide ubiquitous and convenient on-demand access to a shared pool of configurable computing resources. A shared pool of configurable computing resources may be quickly provisioned via virtualization and released with low management effort or service provider interaction, and then scaled accordingly.

[0292] The cloud computing model may consist of various characteristics, such as, for example, on-demand self-service, broad network access, resource pooling, rapid elasticity, metered services, etc. The cloud computing model may also exhibit various service models, such as, for example, software as a service (SaaS), platform as a service (PaaS), and infrastructure as a service (IaaS). The cloud computing model may also be deployed using different deployment models, such as private cloud, community cloud, public cloud, hybrid cloud, etc. In this specification and in the claims, a "cloud computing environment" is an environment in which cloud computing is employed.

[0293] Fig.12 A block diagram of a computing device 1200 that can be configured to perform one or more of the above processes is illustrated. It will be understood that one or more computing devices such as computing device 1200 can implement the detection integration system 106 and the sequencing system 104. Fig.12 As shown, computing device 1200 may include a processor 1202, memory 1204, storage device 1206, I / O interface 1208, and communication interface 1210 that may be communicatively coupled via communication infrastructure 1212. In some embodiments, computing device 1200 may include Fig.12 Those shown may have fewer or more components. The following paragraphs describe in more detail Fig.12 Components of computing device 1200 are shown.

[0294] In one or more embodiments, the processor 1202 includes hardware for executing instructions such as those that constitute a computer program. As an example and not by way of limitation, in order to execute instructions for dynamically modifying a workflow, the processor 1202 may retrieve (or fetch) instructions from an internal register, an internal cache, a memory 1204, or a storage device 1206, and decode and execute them. The memory 1204 may be a volatile or non-volatile memory for storing data, metadata, and programs executed by the processor. The storage device 1206 includes a storage device for storing data or instructions for executing the methods described herein, such as a hard disk, a flash drive, or other digital storage device.

[0295] I / O interface 1208 allows a user to provide input to computing device 1200, receive output from the computing device and otherwise transfer data to the computing device and receive data from the computing device. I / O interface 1208 may include a combination of a mouse, a keypad or keyboard, a touch screen, a camera, an optical scanner, a network interface, a modem, other known I / O devices or such I / O interfaces. I / O interface 1208 may include one or more devices for presenting output to a user, including but not limited to a graphics engine, a display (e.g., a display screen), one or more output drivers (e.g., a display driver), one or more audio speakers and one or more audio drivers. In certain embodiments, I / O interface 1208 is configured to provide graphics data to the display for presentation to the user. Graphics data may represent one or more graphical user interfaces and / or may serve any other graphics content of a specific implementation.

[0296] The communication interface 1210 may include hardware, software, or both. In any case, the communication interface 1210 may provide one or more interfaces for communication (such as, for example, packet-based communication) between the computing device 1200 and one or more other computing devices or networks. As an example and not by way of limitation, the communication interface 1210 may include a network interface controller (NIC) or network adapter for communicating with an Ethernet or other wired-based network, or a wireless NIC (WNIC) or wireless adapter for communicating with a wireless network such as WI-FI.

[0297] Additionally, the communication interface 1210 can facilitate communication with various types of wired or wireless networks. The communication interface 1210 can also facilitate communication using various communication protocols. The communication infrastructure 1212 can also include hardware, software, or both that couple components of the computing device 1200 to each other. For example, the communication interface 1210 can use one or more networks and / or protocols to enable multiple computing devices connected through a particular infrastructure to communicate with each other to perform one or more aspects of the processes described herein. For illustration, a sequencing process can allow multiple devices (e.g., a client device, a sequencing device, and a server device) to exchange information such as sequencing data and error notifications.

[0298] In the foregoing description, the present disclosure has been described with reference to specific exemplary embodiments thereof. Various embodiments and aspects of the present disclosure are described with reference to the details discussed herein, and the accompanying drawings illustrate various embodiments. The above description and drawings are illustrative of the present disclosure and should not be construed as limiting the present disclosure. Many specific details are described to provide a thorough understanding of the various embodiments of the present disclosure.

[0299] The present disclosure may be embodied in other specific forms without departing from its essence or essential features. The embodiments described should be considered as being only exemplary and not restrictive in all respects. For example, the methods described herein may be performed with fewer or more steps / actions, or the steps / actions may be performed in different orders. Additionally, the steps / actions described herein may be repeated or performed in parallel with each other or in parallel with different examples of the same or similar steps / actions. Therefore, the scope of the application is indicated by the appended claims rather than the foregoing description. All changes within the equivalent meaning and scope of the claims will be included within their scope.

Claims

1. A system, comprising: at least one processor; and A non-transitory computer readable medium storing instructions that, when executed by the at least one processor, cause the system to: For one or more genomic coordinates of a genomic sample, receiving a first genotype call for a first type of nucleotide read corresponding to a first threshold number of nucleobases and a second genotype call for a second type of nucleotide read corresponding to a second threshold number of nucleobases; identifying a sequencing metric corresponding to the first genotype call or the second genotype call; generating genotype probabilities for the genotype calls for the one or more genomic coordinates using a genotype call integration machine learning model and based on the sequencing metrics; as well as An output genotype call is generated for the one or more genomic coordinates of the genomic sample based on the genotype probability.

2. The system of claim 1, wherein: The first type of nucleotide reads includes nucleotide reads synthesized from sample library fragments that are shorter than the first threshold number of nucleobases; and The second type of nucleotide reads includes: an assembled nucleotide read, the assembled nucleotide read having been assembled from shorter nucleotide reads to form a continuous sequence satisfying the first threshold number of nucleobases; cycle consensus sequencing (CCS) reads, the CCS reads satisfying the first threshold number of nucleobases; or A nanopore long read, wherein the nanopore long read satisfies the first threshold number of nucleobases.

3. The system of claim 1 , further storing instructions that, when executed by the at least one processor, cause the system to: generating the genotype probabilities by generating genotype probabilities for one or more candidate SNPs using the genotype calling ensemble machine learning model trained with single nucleotide polymorphism (SNP) training data; and The output genotype call is generated indicating the presence or absence of a SNP at the one or more genomic coordinates of the genomic sample.

4. The system of claim 1, further storing instructions that, when executed by the at least one processor, cause the system to generate the output genotype call by: selecting the first genotype detection or the second genotype detection; or A different genotype call is generated that is different from the first genotype call and the second genotype call.

5. The system of claim 1, further storing instructions that, when executed by the at least one processor, cause the system to generate the genotype probabilities by: generating a first genotype probability for the genomic sample comprising a homozygous reference genotype at the one or more genomic coordinates; generating a second genotype probability for the genomic sample comprising a heterozygous variant genotype at the one or more genomic coordinates; as well as A third genotype probability of the genomic sample comprising a homozygous variant genotype at the one or more genomic coordinates is generated.

6. The system of claim 1, wherein the first genotype call comprises a first variant call or a first reference call, and the second genotype call comprises a second variant call or a second reference call.

7. The system of claim 1, wherein the first genotype call or the second genotype call comprises a null data indicator.

8. The system of claim 1, further storing instructions that, when executed by the at least one processor, cause the system to: modifying a genotype metric, a base call quality metric, a genotype quality metric, a genotype probability metric, a genotype likelihood metric, or a PHRED-scaled genotype likelihood metric based on the genotype probability; and A variant call file is generated, the variant call file comprising the modified genotype metric, the modified base call quality metric, the modified genotype quality metric, the modified genotype probability metric, the modified genotype likelihood metric, or the modified PHRED scaled genotype likelihood metric.

9. The system of claim 1, further storing instructions that, when executed by the at least one processor, cause the system to generate the output genotype call by selecting the first genotype call over the second genotype call in the following manner: selecting a homozygous reference genotype call from the first genotype call but not selecting a heterozygous variant genotype call or a homozygous variant genotype call from the second genotype call; selecting the heterozygous variant genotype call from the first genotype call but not selecting the homozygous reference genotype call or the homozygous variant genotype call from the second genotype call; or The homozygous variant genotype call is selected from the first genotype call, but not the heterozygous variant genotype call or the homozygous reference genotype call from the second genotype call.

10. The system of claim 1, further storing instructions that, when executed by the at least one processor, cause the system to generate the output genotype call by selecting the second genotype call instead of the first genotype call in the following manner: selecting a homozygous reference genotype call from the second genotype call instead of selecting a heterozygous variant genotype call or a homozygous variant genotype call from the first genotype call; selecting the heterozygous variant genotype call from the second genotype call instead of selecting the homozygous reference genotype call or the homozygous variant genotype call from the first genotype call; or The homozygous variant genotype call is selected from the second genotype call instead of selecting the heterozygous variant genotype call or the homozygous reference genotype call from the first genotype call.

11. The system of claim 1 , further storing instructions that, when executed by the at least one processor, cause the system to identify the sequencing metric corresponding to the first genotype call or the second genotype call by identifying one or more of: a first set of sequencing metrics associated with the first genotype calls corresponding to nucleotide reads of the first type; a second set of sequencing metrics associated with the second genotype calls corresponding to nucleotide reads of the second type; or A shared set of sequencing metrics associated with both the first genotype call and the second genotype call.

12. The system of claim 1, further storing instructions that, when executed by the at least one processor, cause the system to identify the sequencing metric corresponding to the first genotype call or the second genotype call by determining one or more of: a read-based sequencing metric, a call model-generated sequencing metric, an externally sourced sequencing metric, or a second read type sequencing metric associated with the second genotype call corresponding to the second type of nucleotide read.

13. The system of claim 1, further storing instructions that, when executed by the at least one processor, cause the system to identify the sequencing metric corresponding to the first genotype call or the second genotype call by identifying a read-based sequencing metric, the read-based sequencing metric comprising one or more of: an allele frequency corresponding to an allele of the first genotype call, an allele of the second genotype call, or a different allele of an alternate genotype call different from the first genotype call and the second genotype call; a coverage depth of the first type of nucleotide reads corresponding to the first genotype call or the second type of nucleotide reads corresponding to the second genotype call; an average coverage depth of nucleotide reads of the first type corresponding to the first genotype call or nucleotide reads of the second type corresponding to the second genotype call; a mapping quality metric of the nucleotide reads of the first type corresponding to the first genotype call or the nucleotide reads of the second type corresponding to the second genotype call; or The nucleobase composition of one or more nucleotide reads from the first type of nucleotide reads or the second type of nucleotide reads.

14. The system according to claim 1 further stores instructions, which when executed by the at least one processor cause the system to identify the sequencing metric corresponding to the first genotype call or the second genotype call by identifying the sequencing metric generated by the call model, wherein the sequencing metric generated by the call model includes one or more of the following: a genotype metric of the first genotype call determined based on the first type of nucleotide reads or the second genotype call determined based on the second type of nucleotide reads, a base call quality metric, a genotype quality metric, a genotype probability metric, or a PHRED scaled likelihood metric.

15. The system of claim 1, further storing instructions that, when executed by the at least one processor, cause the system to identify the sequencing metrics corresponding to the first genotype call or the second genotype call by identifying sequencing metrics from an external source, the sequencing metrics from the external source comprising one or more of: a mappability metric indicating difficulty in mapping the nucleotide read to the one or more genomic coordinates within a reference genome; a guanine-cytosine content metric indicative of counts of guanine-cytosine content corresponding to the one or more genomic coordinates within the reference genome; a confidence classification or confidence score indicating the degree to which the nucleobase at the one or more genomic coordinates can be accurately determined; a repeat classification indicating a class of the repeated genomic region for the one or more genomic coordinates; an indicator indicating that the one or more genomic coordinates are part of a cytosine quadruplex (C-quadruplex) within the reference genome; an indicator indicating that the one or more genomic coordinates are part of a guanine quadruplex (G-quadruplex) within the reference genome; or An indicator indicating that the one or more genomic coordinates are part of a homopolymer within the reference genome.

16. The system of claim 1, further storing instructions that, when executed by the at least one processor, cause the system to: receiving the first genotype call by receiving the first genotype call as part of a first variant call file based on nucleotide reads of the first type; receiving the second genotype call by receiving the second genotype call as part of a second variant call file based on nucleotide reads of the second type; and A merged variant call file is generated that includes the first genotype call or the second genotype call.

17. The system of claim 1, further storing instructions that, when executed by the at least one processor, cause the system to: determining that the first genotype call includes a first alternative nucleobase that is different than a second alternative nucleobase of the second genotype call; generating, using the genotype call integrated machine learning model and based on the sequencing metric, a first pipeline accuracy likelihood that the first genotype call is more accurate than the second genotype call and a second pipeline accuracy likelihood that the second genotype call is more accurate than the first genotype call; and The output genotype call is generated by selecting the first genotype call or the second genotype call for the one or more genomic coordinates of the genomic sample based on the first pipeline accuracy likelihood and the second pipeline accuracy likelihood.

18. A system, comprising: at least one processor; and A non-transitory computer readable medium storing instructions that, when executed by the at least one processor, cause the system to: For one or more genomic coordinates of a genomic sample, receiving a first genotype call for a first type of nucleotide read corresponding to a first threshold number of nucleobases and a second genotype call for a second type of nucleotide read corresponding to a second threshold number of nucleobases; identifying a sequencing metric corresponding to the first genotype call or the second genotype call; generating a variant call classification for the candidate variant call at the one or more genomic coordinates using the genotype call ensemble machine learning model and based on the sequencing metrics; as well as An output genotype call is generated for the one or more genomic coordinates of the genomic sample based on the variant call classification.

19. The system of claim 18, further storing instructions that, when executed by the at least one processor, cause the system to: Generating the variant call classification comprises utilizing the genotype calling ensemble machine learning model trained with indel training data to generate the variant call classification for one or more candidate insertions or deletions (indels); and The output genotype call is generated indicating the presence or absence of an indel at the one or more genomic coordinates of the genomic sample.

20. The system of claim 18, further storing instructions that, when executed by the at least one processor, cause the system to generate the variant call classification for the candidate variant by generating one or more of: a first true positive variant probability that the first genotype calls a true positive variant that constitutes the one or more genomic coordinates; a probability that the second genotype calls a second true positive variant that constitutes a true positive variant for the one or more genomic coordinates; The first genotype call comprises a first zygosity error probability of a genotype zygosity error at the one or more genomic coordinates; The second genotype call comprises a second zygosity error probability for a genotype zygosity error at the one or more genomic coordinates; or A reference probability of a homozygous reference genotype at the one or more genomic coordinates.

21. The system of claim 20, further storing instructions that, when executed by the at least one processor, cause the system to: determining that the first true positive variant probability fails to meet a likelihood threshold; and Based on determining that the first true positive variant probability fails to meet the likelihood threshold, the second true positive variant probability is generated or utilized.

22. The system of claim 18, further storing instructions that, when executed by the at least one processor, cause the system to generate the output genotype call by: selecting the first genotype detection or the second genotype detection; or A different genotype call is generated that is different from the first genotype call and the second genotype call.

23. The system of claim 18, wherein: The first type of nucleotide reads includes nucleotide reads synthesized from sample library fragments that are shorter than the first threshold number of nucleobases; and The second type of nucleotide reads includes: an assembled nucleotide read, the assembled nucleotide read having been assembled from shorter nucleotide reads to form a continuous sequence satisfying the first threshold number of nucleobases; cycle consensus sequencing (CCS) reads, the CCS reads satisfying the first threshold number of nucleobases; or A nanopore long read, wherein the nanopore long read satisfies the first threshold number of nucleobases.

24. The system of claim 18, wherein the first genotype call comprises a first variant call or a first reference call, and the second genotype call comprises a second variant call or a second reference call.

25. The system of claim 18, wherein the first genotype call or the second genotype call comprises a null data indicator.

26. The system of claim 18, further storing instructions that, when executed by the at least one processor, cause the system to: modifying a genotype metric, a base call quality metric, a genotype quality metric, a genotype probability metric, a genotype likelihood metric, or a PHRED-scaled genotype likelihood metric based on the variant call classification; and A variant call file is generated, the variant call file comprising the modified genotype metric, the modified base call quality metric, the modified genotype quality metric, the modified genotype probability metric, the modified genotype likelihood metric, or the modified PHRED scaled genotype likelihood metric.

27. The system of claim 18, further storing instructions that, when executed by the at least one processor, cause the system to generate the output genotype call by selecting the first genotype call over the second genotype call in the following manner: selecting a homozygous reference genotype call from the first genotype call but not selecting a heterozygous variant genotype call or a homozygous variant genotype call from the second genotype call; selecting the heterozygous variant genotype call from the first genotype call but not selecting the homozygous reference genotype call or the homozygous variant genotype call from the second genotype call; or The homozygous variant genotype call is selected from the first genotype call, but not the heterozygous variant genotype call or the homozygous reference genotype call from the second genotype call.

28. The system of claim 18, further storing instructions that, when executed by the at least one processor, cause the system to generate the output genotype call by selecting the second genotype call instead of the first genotype call in the following manner: selecting a homozygous reference genotype call from the second genotype call instead of selecting a heterozygous variant genotype call or a homozygous variant genotype call from the first genotype call; selecting the heterozygous variant genotype call from the second genotype call instead of selecting the homozygous reference genotype call or the homozygous variant genotype call from the first genotype call; or The homozygous variant genotype call is selected from the second genotype call instead of selecting the heterozygous variant genotype call or the homozygous reference genotype call from the first genotype call.

29. The system of claim 18, further storing instructions that, when executed by the at least one processor, cause the system to identify the sequencing metric corresponding to the first genotype call or the second genotype call by identifying one or more of: a first set of sequencing metrics associated with the first genotype calls corresponding to nucleotide reads of the first type; a second set of sequencing metrics associated with the second genotype calls corresponding to nucleotide reads of the second type; or A shared set of sequencing metrics associated with both the first genotype call and the second genotype call.

30. The system of claim 18, further storing instructions that, when executed by the at least one processor, cause the system to identify the sequencing metric corresponding to the first genotype call or the second genotype call by determining one or more of: a read-based sequencing metric, a call model-generated sequencing metric, an externally sourced sequencing metric, or a second read type sequencing metric associated with the second genotype call corresponding to the second type of nucleotide read.

31. The system of claim 18, further storing instructions that, when executed by the at least one processor, cause the system to identify the sequencing metric corresponding to the first genotype call or the second genotype call by identifying a read-based sequencing metric, the read-based sequencing metric comprising one or more of: an allele frequency corresponding to an allele of the first genotype call, an allele of the second genotype call, or a different allele of an alternate genotype call different from the first genotype call and the second genotype call; a coverage depth of the first type of nucleotide reads corresponding to the first genotype call or the second type of nucleotide reads corresponding to the second genotype call; an average coverage depth of nucleotide reads of the first type corresponding to the first genotype call or nucleotide reads of the second type corresponding to the second genotype call; a mapping quality metric of the nucleotide reads of the first type corresponding to the first genotype call or the nucleotide reads of the second type corresponding to the second genotype call; or The nucleobase composition of one or more nucleotide reads from the first type of nucleotide reads or the second type of nucleotide reads.

32. The system of claim 18, further storing instructions that, when executed by the at least one processor, cause the system to identify the sequencing metric corresponding to the first genotype call or the second genotype call by identifying the sequencing metric generated by a call model, wherein the sequencing metric generated by the call model includes one or more of the following: a genotype metric for the first genotype call determined based on the first type of nucleotide reads or the second genotype call determined based on the second type of nucleotide reads, a base call quality metric, a genotype quality metric, a genotype probability metric, a genotype likelihood metric, or a PHRED scaled likelihood metric.

33. The system of claim 18, further storing instructions that, when executed by the at least one processor, cause the system to identify the sequencing metrics corresponding to the first genotype call or the second genotype call by identifying sequencing metrics from an external source, the sequencing metrics from the external source comprising one or more of: a mappability metric indicating difficulty in mapping the nucleotide read to the one or more genomic coordinates within a reference genome; a guanine-cytosine content metric indicative of counts of guanine-cytosine content corresponding to the one or more genomic coordinates within the reference genome; a confidence classification or confidence score indicating the degree to which the nucleobase at the one or more genomic coordinates can be accurately determined; a repeat classification indicating a class of the repeated genomic region for the one or more genomic coordinates; an indicator indicating that the one or more genomic coordinates are part of a cytosine quadruplex (C-quadruplex) within the reference genome; an indicator indicating that the one or more genomic coordinates are part of a guanine quadruplex (G-quadruplex) within the reference genome; or An indicator indicating that the one or more genomic coordinates are part of a homopolymer within the reference genome.

34. The system of claim 18, further storing instructions that, when executed by the at least one processor, cause the system to: receiving the first genotype call by receiving the first genotype call as part of a first variant call file based on nucleotide reads of the first type; receiving the second genotype call by receiving the second genotype call as part of a second variant call file based on nucleotide reads of the second type; and A merged variant call file is generated that includes the first genotype call or the second genotype call.

35. The system of claim 18, further storing instructions that, when executed by the at least one processor, cause the system to: determining that the first genotype call includes a first alternative nucleobase that is different than a second alternative nucleobase of the second genotype call; generating, using the genotype call integrated machine learning model and based on the sequencing metric, a first pipeline accuracy likelihood that the first genotype call is more accurate than the second genotype call and a second pipeline accuracy likelihood that the second genotype call is more accurate than the first genotype call; and The output genotype call is generated by selecting the first genotype call or the second genotype call for the one or more genomic coordinates of the genomic sample based on the first pipeline accuracy likelihood and the second pipeline accuracy likelihood.

Images