A method and apparatus for discriminating sample contamination in high-throughput sequencing

Abstract

The present disclosure relates to the technical field of sequencing data processing, and particularly relates to a method and device for identifying sample contamination in high-throughput sequencing. The present disclosure provides a method for constructing a feature iSNP database for identifying sample contamination in high-throughput sequencing, which comprises the following steps: S1, selecting a standard alignment genome of a target strain; S2, obtaining reference genomes of at least two strains of the target strain, aligning the reference genomes of the strains with the standard alignment genome, obtaining specific SNP sites of each reference genome, and constructing a basic iSNP database; S3, aligning sequencing data of a preset sample with the standard alignment genome, obtaining SNP sites of the preset sample relative to the standard alignment genome, and constructing a feature iSNP database according to the SNPs of the preset sample. The present disclosure can realize low-cost and high-precision detection of contamination in high-throughput sequencing samples.

Description

Technical Field

[0001] This invention relates to the field of sequencing data processing technology, and in particular to a method and apparatus for identifying sample contamination in high-throughput sequencing. Background Technology

[0002] Accurate identification of pathogenic microorganisms is a core aspect of clinical diagnosis. With advancements in sequencing technology, next-generation sequencing (NGS) has gradually replaced traditional culture and PCR techniques, becoming a crucial tool for microbial detection. Metagenomic sequencing (mNGS), in particular, performs unbiased sequencing of all nucleic acids (including DNA and RNA) in a sample without the need for pre-defined targets, enabling the simultaneous detection of multiple pathogens, including viruses, bacteria, fungi, and parasites. This comprehensive approach demonstrates significant potential in scenarios such as nervous system infections, bloodstream infections, and tracing the origins of emerging epidemics. To enhance the detection sensitivity of specific pathogens, targeted NGS (tNGS) technology has emerged. tNGS is divided into two technical routes: multiplex tNGS and probe-targeted capture tNGS. While both NGS technologies offer higher sensitivity than traditional methods, contamination during the experimental process becomes more prominent, posing a significant obstacle to accurate diagnosis. Therefore, establishing a contamination identification system is essential for the clinical application of pathogen NGS.

[0003] Throughout the entire process of pathogen NGS detection, contamination is prevalent and originates from various sources, severely impacting the reliability of results. The most prominent issue is background microbial contamination of reagents. Since reagent production may introduce nucleic acids from engineered bacteria and environmental microorganisms, these contaminated sequences can lead to false positives in low-biomass samples (such as blood and cerebrospinal fluid). Aerosol contamination from the same batch of samples is another major challenge. During nucleic acid extraction, PCR amplification, or library construction, aerosols can cause inter-sample contamination. For example, when a high-throughput sample processing plate contains both strongly positive samples and negative controls, aerosols may carry pathogen nucleic acids and diffuse to adjacent wells. This contamination is extremely common in high-throughput operations using 96-well or 384-well plates and is difficult to completely eliminate through routine disinfection. Furthermore, index hopping is a specific problem in dual-indexed sequencing. When sequencing clusters are misread in the flow cell due to fluorescence signals, leading to incorrect index allocation, reads from one sample can be misassigned to another, creating a false "low-level infection" signal. Residual contamination from historical samples is closely related to experimental equipment. For example, if the channels of automated pipettes, nucleic acid extractors, or sequencers are not thoroughly cleaned, nucleic acid fragments from previous samples may remain. These residual DNA fragments may be amplified and detected in subsequent tests. The complexity of contamination lies in its dynamic nature and cumulative effect. Factors such as temperature, humidity, and reagent batches can all affect the level of contamination. Furthermore, contamination can gradually accumulate at multiple stages, including nucleic acid extraction, library construction, and sequencing, leading to amplified contamination signals in the final data.

[0004] Currently, mainstream solutions for addressing contamination issues in NGS detection of pathogens have significant limitations. For example, setting up negative controls (such as NTC) can only indicate the presence of contamination, but cannot accurately trace the specific source of contamination (e.g., it cannot distinguish between reagent-engineered bacteria residues and environmental microorganisms introduced during experimental procedures), and its sensitivity in identifying low-abundance contamination is insufficient. In addition, systems such as the MIST system developed by Song et al. (Imeta. 2023 Nov 2;2(4):e146) use phylogenetic tree construction for microbial origin tracing, but their method heavily relies on complete genomic data, and its generalization ability and practicality are significantly reduced when faced with clinical samples with insufficient NGS sequencing depth and incomplete genomic coverage. For cross-contamination between samples, although the approach of tracing origins based on SNP allele frequency deviation in tumor ctDNA detection can be referenced, the high diversity of microbial genomes makes universal SNP site screening extremely difficult, and the depth of pathogens in NGS detection of clinical samples is usually limited. How to eliminate false positive SNPs caused by PCR errors and sequencing errors is another challenge.

[0005] In summary, existing technologies lack a high-precision, high-sensitivity contamination identification scheme that can systematically integrate multi-source contamination characteristics and is applicable to the characteristics of clinical NGS data. Summary of the Invention

[0006] To overcome the aforementioned problems in the existing technology, this disclosure provides a method for constructing a characteristic iSNP database for identifying sample contamination in high-throughput sequencing, a method for identifying sample contamination in high-throughput sequencing, and an apparatus. This method and apparatus for identifying sample contamination in high-throughput sequencing can effectively identify reagent background microorganisms and inter-sample contamination, reducing false negatives and false positives, and has significant implications for the identification of clinical pathogenic microorganisms.

[0007] According to a first aspect of this disclosure, a method for constructing a characteristic iSNP database for identifying sample contamination in high-throughput sequencing is provided, the method comprising the following steps: S1, Select the standard alignment genome of the target bacterial species; S2, obtain reference genomes of at least two strains of the target bacterial species, compare the reference genomes of the strains with the standard alignment genome, obtain specific SNP sites for each reference genome, and construct a basic iSNP database; S3, compare the sequencing data of the preset sample with the standard aligned genome to obtain the SNP sites of the preset sample relative to the standard aligned genome, and construct a feature iSNP database based on the SNP sites of the preset sample.

[0008] In some implementations, the target bacterial species are one or more bacterial species that cause sample contamination in high-throughput sequencing.

[0009] In some implementations, step S2, obtaining each reference genome-specific SNP site and constructing a basic iSNP database, includes: SNP sites that differ from other reference genomes in at least one reference genome are retained, while duplicated SNP sites and SNP sites located in highly duplicated regions are removed.

[0010] In some implementations, when the preset sample is a negative control sample, step S2, obtaining each reference genome-specific SNP locus and constructing a basic iSNP database, further includes grouping the reference genome and calculating the specificity of each SNP locus within the group using Formula 1. ,Will SNPs with values ​​≥ a first preset threshold are defined as iSNPs for that group, thus obtaining the basic iSNP database. Formula 1 in, The frequency of the SNP in all reference genomes in group i is represented by , and n represents the number of groups. In a preferred embodiment, the first preset threshold is 0.6-1.0, for example, 0.6, 0.7, 0.8, 0.9 or 1.0.

[0011] In a preferred embodiment, grouping the reference genome may include grouping based on clinical or biological significance, such as grouping according to the phylogenetic relationship of the strains, the pathogenicity of the strains, or the origin of isolation. In a more preferred embodiment, the reference genome is grouped according to the phylogenetic relationship of the strains.

[0012] In some implementations, in step S3, the preset sample may be selected from negative control samples or suspected contamination source samples.

[0013] In some implementations, in step S3, when the preset sample is a negative quality control sample, constructing a feature iSNP database based on the SNP sites of the preset sample includes: SNP sites detected in at least 50%, 60%, 70%, 80%, 90%, and 100% of the negative control samples were retained to obtain the reagent background strain iSNP. In some implementations, in step S3, when the preset sample is a suspected source of contamination, constructing a feature iSNP database based on the SNP sites of the preset sample includes: The SNP sites of suspected pollution source samples are compared with the basic iSNP database. SNP sites that match the basic iSNPs and have a stable frequency in the suspected pollution source samples are identified as suspected pollution source characteristic iSNPs.

[0014] In some implementations, the SNP site with stable frequency refers to an SNP site whose allele frequency is in the range of 0.8-1.0, for example: 0.8, 0.9, 1.0.

[0015] In some embodiments, the characteristic iSNP database includes reagent background strain iSNPs, a basic iSNP library, and / or characteristic iSNPs of suspected contaminants. In some embodiments, the contamination includes reagent background strain contamination and inter-sample contamination.

[0016] In some implementations, in step S1, the screening criteria for the standard alignment genome include: (1) Sourced from databases that have undergone manual review, including but not limited to NCBI RefSeq; (2) Select according to the priority of assembly level: Completed diagram > Chromosome level > Scaffold > Contig. Preferably, the assembly level is the completed diagram or chromosome level; (3) The integrity of the genome is ≥95% and the contamination level is ≤5%.

[0017] In some embodiments, the high-throughput sequencing includes high-throughput sequencing of pathogenic microorganisms.

[0018] According to another aspect of this disclosure, a method for identifying sample contamination in high-throughput sequencing is provided, the method comprising the following steps: A1, Obtain the SNP sites of the sample to be identified; and A2. The SNP sites obtained in step A1 are compared with the characteristic iSNP database obtained by the method described in this disclosure to determine the source of the microorganisms in the sample to be identified.

[0019] In some embodiments, when the contamination is background contamination from reagent strains, step A2 includes the following steps: B1. The SNP sites obtained in step A1 are compared with the characteristic iSNP database obtained by the method described in this disclosure, and the sequencing sequences containing the SNP sites obtained in step A1 are classified. B2. After all sequences are annotated, the number of sequencing sequences from all sources is counted, and the confidence level M is calculated. If M ≥ a second preset threshold, the microorganisms in the sample to be identified are classified into the group with the most sequences. Formula 2 in, This represents the sequence number of the group with the most sequences. Let i represent the sequence number of group i, and n represent the number of groups.

[0020] In a preferred embodiment, the second preset threshold is 0.6-1.0, for example, 0.6, 0.7, 0.8, 0.9, or 1.0.

[0021] In some implementations, in step B1, the classification of the sequencing sequences containing the SNP sites obtained in step A1 includes the following criteria: (b1) If a sequencing sequence of a sample to be identified contains at least two reagent background strain iSNPs, then the sequencing sequence is classified as reagent background. (b2) If a sequencing sequence of a sample to be identified contains at least two iSNPs of the same group and no iSNPs of other groups, then the sequencing sequence is classified into the corresponding group. (b3) When a sequencing sequence of a sample to be identified contains at least two iSNPs of the same group and at least two iSNPs of the reagent background strain, the sequencing sequence is classified as the reagent background.

[0022] In some implementations, when the contamination is inter-sample contamination, step A2 includes: The SNP sites obtained in step A1 are compared with the feature iSNP database obtained by the method described in this disclosure, and the similarity between the SNP spectrum of the sample to be identified and the feature iSNP spectrum of the suspected pollution source is calculated. The true source of pollution is determined based on the similarity.

[0023] In some implementations, the similarity Calculated according to Formula 3 Formula 3 in, This indicates the number of all SNPs in the sample to be identified that share the same iSNP location as the suspected source of pollution. This indicates the number of SNPs in the sample to be identified that have the same bases as the iSNP sites characteristic of the suspected pollution source.

[0024] In some implementations, the criteria for determining the true source of pollution based on similarity are as follows: when 60%, 70%, 80%, 90%, or 100% of the samples to be identified have a C value greater than or equal to a third preset threshold, batch contamination is considered to have occurred, and the suspected source of pollution is the source of pollution. In a preferred implementation, the third preset threshold is 0.6, 0.7, 0.8, 0.9, or 1.0.

[0025] In some embodiments, when the contamination is inter-sample contamination, the method further includes: A3 filters out contaminated sequences to determine whether they represent a true infection.

[0026] In some implementations, filtering the contaminating sequence in step A3 includes: D1, for each contaminated sample other than the contamination source, a sequence carrying ≥2 characteristic iSNPs of the identified contamination source on one of its sequencing sequences is marked as a contaminated sequence and filtered to obtain the filtered sequence; and D2, when the number of filtered sequences is ≥10, the sample to be identified is a true infection.

[0027] According to another aspect of this disclosure, an apparatus for identifying sample contamination in high-throughput sequencing is provided, the apparatus comprising: Sequencing information acquisition module: configured to acquire sequencing information obtained after high-throughput sequencing of the sample to be tested; Sample status determination module: configured to determine the sample contamination status according to the method for identifying sample contamination in high-throughput sequencing obtained by the construction method described in this disclosure, based on the iSNP feature database for identifying sample contamination in high-throughput sequencing.

[0028] In this application, S1, S2, S3, etc. are used only to identify steps and do not represent a necessary execution order between steps. Detailed Implementation

[0029] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to embodiments. The specific embodiments described herein are for illustrative purposes only and are not intended to limit the invention in any way. Furthermore, descriptions of well-known structures and techniques are omitted in the following description to avoid unnecessarily obscuring the concepts of this disclosure. Such structures and techniques have also been described in many publications.

[0030] definition Unless otherwise defined, all technical and scientific terms used in this invention have the same meaning as commonly used in the field to which this invention pertains. For the purposes of interpreting this specification, the following definitions will apply, and where appropriate, terms used in the singular will also include the plural forms, and vice versa.

[0031] Unless the context clearly indicates otherwise, the terms “a” and “an” as used herein include plural references. For example, reference to “a cell” includes multiple such cells and equivalents known to those skilled in the art, etc.

[0032] As used herein, the term "about" indicates a range of ±20% of the following value. In some embodiments, the term "about" indicates a range of ±10% of the following value. In some embodiments, the term "about" indicates a range of ±5% of the following value.

[0033] The following examples are provided to aid in understanding the invention. However, it should be understood that these examples are for illustrative purposes only and do not constitute any limitation. The actual scope of protection of the invention is set forth in the claims. It should be understood that any modifications and changes can be made without departing from the spirit of the invention. The reagents and / or kits used in the following examples are commercially available or can be synthesized by known methods.

[0034] It should be noted that, unless specific conditions are specified in the examples, experimental conditions should be performed according to standard conditions, manufacturer recommendations, or publicly reported experimental conditions. Reagents or instruments whose manufacturers are not specified are all commercially available, standard products. For reagents whose manufacturers are specified, similar products from other manufacturers are substitutes.

[0035] Unless otherwise specified, the quantitative experiments in the following examples were all repeated three times, and the results were averaged.

[0036] Example Example 1: Escherichia coli Reagent Background Identification and Filtering Escherichia coli, as one of the most commonly used engineered strains in molecular biology, is one of the most frequent sources of reagent background contamination in microbial NGS assays. Currently, most commonly used commercial E. coli strains are derived from two isolates: K-12 and B strains. For example, DH10B and DH5α strains, commonly used for cloning vector amplification, are K-12 derivatives, while BL21(DE3), commonly used for protein expression, is a common strain within the B strain. During the culture of engineered strains, plasmid extraction, or enzyme purification, genomic DNA or plasmid DNA fragments may remain in commercial reagents. When testing low-biomass clinical samples (such as cerebrospinal fluid and blood), these contaminants are simultaneously amplified and sequenced, producing false positive signals and severely interfering with clinical judgment. Particularly challenging is that some pathogenic E. coli strains are completely identical in species classification to engineered strains, making it impossible to distinguish between infection and contamination based solely on species name. This can lead to unnecessary antibiotic treatment or incorrect infection tracing. Therefore, identifying and filtering E. coli reagent background is crucial for improving the specificity of clinical NGS assays.

[0037] 1. Construction of the Escherichia coli iSNP database 1.1 Selection of Standards for Genome Alignment The *Escherichia coli* subspecies K-12 MG1655 (Assembly: GCF_000005845.2) was selected from the NCBI RefSeq database as the standard genome for alignment. This genome is the most thoroughly studied and highly complete reference sequence, serving as the benchmark for all subsequent alignments to ensure the consistency of SNP coordinates and the reproducibility of the analysis.

[0038] 1.2 Constructing a basic iSNP database based on a public reference genome a. Data collection: Download all reference genomes labeled "Escherichia coli" and assembled at the "Complete Genome" or "Chromosome" level from the RefSeq database.

[0039] b. Quality control and screening: The downloaded genomes were evaluated using CheckM software to screen out high-quality genomes with integrity ≥95% and contamination ≤5%.

[0040] c. Whole Genome Alignment and SNP Identification: The nucmer command in the MUMmer software package was used to perform a whole genome alignment of each high-quality reference genome with the standard aligned genome. Subsequently, the show-snps command was used to accurately identify the SNP sites of each strain relative to the standard aligned genome.

[0041] d. SNP Annotation and Filtering: Identified SNPs were filtered, retaining sites that differed from other strains in at least one strain, and removing duplicates and sites located in highly repetitive regions, resulting in a basic SNP set containing thousands of polymorphic sites across different Escherichia coli strains. Then, based on the reference (Mash-based analyses of Escherichia coli genomes reveal 14 distinct phylogroups, DOI: 10.1038 / s42003-020-01626-5), the Escherichia coli genome was divided into 10 distinct groups. Groups B2-1 and B2-2 were merged into group B2, and groups D1, D2, and D3 were merged into group D. The specificity of each SNP within each group was calculated. S That is, for each SNP locus, its frequency of occurrence in each predefined group is calculated, and specificity indicators are used. S The quantification and calculation formula is as follows: Formula 1 in, This indicates the frequency of the SNP site in all reference genomes in group i, and n represents the number of groups.

[0042] Will S SNPs with a value ≥ 0.8 are defined as informative SNPs (iSNPs) for that group and are used for subsequent precise tracing.

[0043] 1.3 Constructing an iSNP database of "reagent background strains" based on historical negative reference materials a. Data collection: Collect NGS sequencing data of negative controls included in all test batches of Geneplus Medical Laboratory in the past year.

[0044] b. Data analysis: Standard bioinformatics analysis was performed on these negative control data. After quality control and human-free sequencing, classification tools such as Kraken2 were used to confirm the presence of Escherichia coli sequences.

[0045] c. Target sequence alignment and SNP identification: The Escherichia coli sequence identified in the negative control was aligned with the standard alignment genome (MG1655 strain) using BWA-MEM. SNP identification was performed using Samtools with strict filtering conditions set (e.g., base quality ≥30, alignment quality ≥20).

[0046] d. Feature SNP extraction: SNP sites that were detected in more than 60% of historical negative controls were identified as SNP sites specifically enriched in the negative controls and defined as "reagent background strain" iSNPs. The number of specific iSNPs in each group is shown in Table 1.

[0047] Table 1. Number of iSNPs in each group of Escherichia coli

[0048] 2. Background identification of clinical test sample reagents 2.1 Clinical sample testing and target sequence identification Twenty-three samples were collected for testing, including 13 clinically submitted blood culture-positive blood samples (samples 1-13) and 10 negative control samples (NTCs, samples 14-23) tested concurrently. All samples underwent NGS sequencing, and after quality control and dehumanization, microbial classification was performed using Kraken2 to confirm the presence of Escherichia coli sequences in all samples.

[0049] 2.2 Target Sequence Alignment and SNP Identification The Escherichia coli sequences identified in each sample were compared with the standard alignment genome (MG1655) using BWA-MEM. SNPs were identified using tools such as GATK or Samtools, and the following filtering conditions were set: base quality ≥30, alignment quality ≥20.

[0050] 2.3 Annotate sample SNPs using the iSNP feature database The SNP sites obtained for each sample are compared with the characteristic iSNP database constructed in step 1 of this embodiment. For each sample, the number of iSNPs that can be annotated by the database in its SNP sites is counted, and the main groups to which it is annotated (such as group B2, group D, reagent background strain, etc.) are recorded.

[0051] Because the sequences of target microorganisms in clinical samples are typically low-resolution and high-depth SNPs are rare, linkage analysis is performed on the SNPs of each sample to reduce interference from PCR or sequencing errors. Specifically, a sequence in each sample is classified if it contains at least two iSNPs; otherwise, the sequence is discarded. The details are as follows: First, the SNP sites of the test sample are compared with the characteristic iSNP database to calculate the number of matching sequences: when a sequencing sequence of the test sample contains at least two "non-reagent background strain" iSNPs of the same group, the sequence is considered to be from clinical infection; when a sequencing sequence of the test sample contains at least two reagent background strain iSNPs, the sequence is considered to be from reagent background, otherwise it is considered to be introduced by PCR or sequencing error; when a sequencing sequence of the test sample contains at least two "non-reagent background strain" iSNPs of the same group and at least two reagent background strain iSNPs, the sequence is classified as reagent background.

[0052] Secondly, sequences from sources other than the reagent background are annotated: if a sequence contains two or more iSNPs from the same group and no iSNPs from other groups, the sequence is annotated as originating from that group; otherwise, it is marked as "other". After all sequences are annotated, the number of sequences from all annotated sources is counted, and the confidence level M is calculated using the following formula. If M ≥ 0.8, then the species in this sample belongs to the group with the most sequences from a clearly identified source. Formula 2 in, This represents the sequence number of the group with the most sequences. Let i represent the sequence number of group i, and n represent the number of groups.

[0053] After linkage analysis, the number of sequences containing non-reagent background strain linkage iSNPs in the sample and the main groups annotated with the linkage iSNPs in the sample were recorded. The results are shown in Table 2.

[0054] Table 2 Results of iSNP analysis of Escherichia coli in clinical samples

[0055] Table 3 shows that the number of sequences detected by NGS in blood culture-positive clinical samples has a wide distribution range, with no significant difference between the number of sequences detected in samples 9-13 and the negative control. After iSNP analysis to exclude sequences from reagent background, the number of sequences in positive samples was all above 10, while the number of sequences in negative controls was no more than 2, indicating that iSNP can enhance the ability to distinguish reagent background during NGS detection and improve the detection sensitivity of low-load clinical samples. In addition, sample 12 had two or more identical sequence numbers in two groups, which could not be annotated to the same group, but the sample could still be identified as positive. Furthermore, 12 of the 13 positive samples were successfully annotated to known Escherichia coli groups, with group B2 appearing most frequently. This is consistent with the isolation rate of Escherichia coli in bloodstream infections reported in the literature (Mash-based analysesof Escherichia coli genomes reveal 14 distinct phylogroups, DOI: 10.12122 / j.issn.1673-4254.2018.12.20).

[0056] The above results fully demonstrate that the method disclosed herein can effectively eliminate reagent background interference and greatly reduce the risk of false positives; moreover, the method can retain and confirm the true low-volume infection signal in strong background noise, avoid false negatives caused by over-filtering, and significantly improve the accuracy and reliability of clinical low-biomass sample detection.

[0057] Example 2: Inter-sample contamination analysis for pathogen detection In NGS detection of pathogenic microorganisms, samples with extremely high viral loads are highly susceptible to contamination of other samples in the same batch via aerosols or indexhopping, leading to false positive results and severely interfering with clinical decision-making. Traditional methods cannot effectively distinguish such contamination from true low-level infection. This embodiment uses human herpesvirus type 5 (CMV) as an example to verify the ability of the method disclosed herein to accurately identify contamination between samples, trace the source of contamination, and reconstruct the true signal.

[0058] 1. Construction of the SNP Feature Database The CMV Merlin standard strain (Accession: NC_006273.2) was selected from the NCBI RefSeq database as the standard genome for alignment. This genome is a complete reference sequence, ensuring the consistency of SNP coordinates. Clinical isolate genome sequences of CMV were downloaded from public databases (e.g., NCBI Refseq, NCBI GenBank, ENA database), and quality control screening was performed. Using tools such as the MUMmer software package, each isolate was aligned to the Merlin standard strain to identify SNP sites (the method is consistent with step ac in section 1.2 of Example 1). The identified SNPs were filtered, retaining sites that differed from other isolates in at least one isolate, and removing duplicates and sites located in highly repetitive regions. The remaining sites were annotated with reference genome sources to form the basic iSNP database.

[0059] NGS data of samples identified as suspected CMV sources in historical testing by Geneplus Medical Laboratory were included in the analysis. In this example, CMV was detected in all 16 blood samples from the same batch, and their sequence numbers are shown in Table 3. Samples 1 and 2 in this batch had the highest CMV load, potentially contaminating other samples. Therefore, Sample 1 was designated as suspected source 1, and Sample 2 as suspected source 2. The sequences of Samples 1 and 2 were compared with the standard aligned genome of the Merlin standard strain. Following step 1.3c in Example 1, SNP sites in Samples 1 and 2 were identified. Sites matching the basic iSNP database and exhibiting stable frequencies in the sample were defined as characteristic iSNPs of the suspected source of contamination. This resulted in 2584 characteristic iSNPs for Sample 1 and 1197 characteristic iSNPs for Sample 2. Stable frequency sites were defined as SNPs with an allele frequency ≥ 0.8.

[0060] Table 3. Number of sequences detected in samples from potentially contaminated batches.

[0061] 2. Source tracing of pollution NGS data were collected from the 16 clinical plasma samples mentioned above. The identified CMV sequences in each sample were aligned with standard aligned genomes to obtain the SNP loci for each sample. The similarity between the SNP profile of each sample and the characteristic iSNP profiles of the two suspected contaminants, Sample 1 and Sample 2, was calculated. C Samples with higher similarity are more likely to be suspected sources of contamination. Similarity C The calculation formula is as follows: when more than 60% of the samples in a batch... C When the concentration is ≥0.8, batch contamination can be considered to have occurred, and the strongly positive sample is the source of contamination. Formula 3 in, This indicates the number of all SNPs in the sample that share the same location as the iSNP site characteristic of the suspected source of contamination. This indicates the number of SNPs in the sample that share the same bases as the characteristic iSNP sites of a suspected pollution source. An SNP spectrum is the set of all SNP sites in a sample, while a characteristic iSNP spectrum is the set of all characteristic iSNP sites in a suspected pollution source sample.

[0062] The results are shown in Table 4. It can be seen that 12 out of the 14 samples are more similar to sample 1, indicating that this batch of samples is very likely to have been contaminated by sample 1.

[0063] Table 4. Similarity to suspected pollution sources C Statistical results

[0064] 3. Batch contamination filtration Using sample 1 as the contamination source, linkage iSNP analysis was used to filter contaminating sequences in other samples. For each contaminated sample, sequences carrying ≥2 iSNPs that matched the identified contamination source characteristics were marked as "contaminating sequences" and filtered. After filtering, the number of remaining CMV sequences in each sample that did not belong to any contamination source was counted, and the results are shown in Table 5.

[0065] Table 5. Pollution filtration results with Sample 1 as the pollution source.

[0066] As shown in Table 5, after filtering, the CMV sequence counts of all qPCR-negative samples (6-16) decreased to extremely low levels (≤4) or 0, which is consistent with their clinical expectations and effectively avoids false positive reports. After filtering, the qPCR-positive samples (2,3,4,5) all retained more than 10 CMV sequences, corroborating their clinical evidence of qPCR positivity. These results demonstrate that the method disclosed in this paper can not only remove inter-sample contamination but also retain and confirm true low-level infection in severe cross-contamination backgrounds, avoiding false negatives caused by contamination masking. This verifies that the method disclosed in this paper possesses the dual advantages of high-specificity decontamination and high-sensitivity fidelity in complex contamination backgrounds.

[0067] The technical solutions of the present invention are not limited to the specific embodiments described above. Any technical modifications made in accordance with the technical solutions of the present invention fall within the protection scope of the present invention.

Claims

1. A method for constructing a characteristic iSNP database for identifying sample contamination in high-throughput sequencing, characterized in that, The method includes the following steps: S1, Select the standard alignment genome of the target bacterial species; S2, obtain reference genomes of at least two strains of the target bacterial species, align the reference genomes of the strains with the standard aligned genome, obtain specific SNP sites for each reference genome, and construct a basic iSNP database; and S3, compare the sequencing data of the preset sample with the standard aligned genome to obtain the SNP sites of the preset sample relative to the standard aligned genome, and construct a feature iSNP database based on the SNP sites of the preset sample.

2. The construction method according to claim 1, characterized in that, In step S2, obtaining each reference genome-specific SNP site and constructing the basic iSNP database includes: SNP sites that differ from other reference genomes in at least one reference genome are retained, while duplicated SNP sites and SNP sites located in highly duplicated regions are removed.

3. The construction method according to claim 2, characterized in that, When the preset sample is a negative quality control sample, step S2, in which the specific SNP loci for each reference genome are obtained and the basic iSNP database is constructed, further includes grouping the reference genome and calculating the specificity of each SNP locus within the group using Formula 1. ,Will SNPs with values ​​≥ a first preset threshold are defined as iSNPs for that group, thus obtaining the basic iSNP database. Official 1 in, This represents the frequency of the SNP across all reference genomes in group i, where n represents the number of groups. Preferably, the first preset threshold is 0.6-1.

0. Preferably, grouping the reference genome includes grouping based on clinical or biological significance. More preferably, the reference genome is grouped according to the phylogenetic relationship of the strains, the pathogenicity of the strains, or the source of isolation. More preferably, the reference genome is grouped according to the phylogenetic relationship of the strains.

4. The construction method according to claim 1, characterized in that, In step S3, when the preset sample is a negative quality control sample, constructing a feature iSNP database based on the SNP sites of the preset sample includes: Retain at least 50%, 60%, 70%, 80%, 90%, and 100% of the negative control samples that were detected to obtain the reagent background strain iSNP.

5. The construction method according to claim 1, characterized in that, In step S3, when the preset sample is a suspected source of contamination, constructing a feature iSNP database based on the SNP sites of the preset sample includes: The SNP sites of suspected pollution source samples are compared with the basic iSNP database. SNP sites that match the basic iSNPs and have a stable frequency in the suspected pollution source samples are identified as suspected pollution source characteristic iSNPs. Preferably, the SNP site with stable frequency refers to an SNP site whose allele frequency is in the range of 0.8-1.

0.

6. The construction method according to claim 1, characterized in that, In step S1, the screening criteria for the standard alignment genome include: (1) It comes from a database that has been manually reviewed and edited, preferably from the NCBI RefSeq database; (2) Select according to the priority of assembly level: Completed diagram > Chromosome level > Scaffold > Contig. Preferably, the assembly level is the completed diagram or chromosome level; (3) The integrity of the genome is ≥95% and the contamination level is ≤5%.

7. A method for identifying sample contamination in high-throughput sequencing, characterized in that, The method includes the following steps: A1, Obtain the SNP sites of the sample to be identified; and A2. Compare the SNP sites obtained in step A1 with the characteristic iSNP database obtained by the construction method described in any one of claims 1-6 to determine the source of the microorganisms in the sample to be identified.

8. The method according to claim 7, characterized in that, When the contamination is background contamination from the reagent, step A2 includes the following steps: B1. The SNP sites obtained in step A1 are compared with the feature iSNP database obtained by the construction method described in any one of claims 1-6, and the sequencing sequences containing the SNP sites obtained in step A1 are classified. B2. After all sequences are annotated, the number of sequencing sequences from all sources is counted, and the confidence level M is calculated. If M ≥ a second preset threshold, the microorganisms in the sample to be identified are classified into the group with the most sequences. Official 2 in, This represents the sequence number of the group with the most sequences. Let represent the sequence number of group i, and n represent the number of groups. Preferably, the second preset threshold is 0.6-1.

0. Preferably, in step B1, the classification of the sequencing sequences containing the SNP sites obtained in step A1 includes the following criteria: (b1) If a sequencing sequence of a sample to be identified contains at least two reagent background strain iSNPs, then the sequencing sequence is classified as reagent background. (b2) If a sequencing sequence of a sample to be identified contains at least two iSNPs of the same group and no iSNPs of other groups, then the sequencing sequence is classified into the corresponding group. (b3) When a sequencing sequence of a sample to be identified contains at least two iSNPs of the same group and at least two iSNPs of the reagent background strain, the sequencing sequence is classified as the reagent background.

9. The method according to claim 7, characterized in that, When the contamination is inter-sample contamination, step A2 includes: The SNP sites obtained in step A1 are compared with the feature iSNP database obtained by the method described in any one of claims 1-6, and the similarity between the SNP spectrum of the sample to be identified and the feature iSNP spectrum of the suspected pollution source is calculated. The true source of pollution is determined based on similarity. Preferably, the similarity Calculated according to Formula 3 Official 3 in, This indicates the number of all SNPs in the sample to be identified that share the same iSNP location as the suspected source of pollution. This indicates the number of SNPs in the sample to be identified that have the same bases as the iSNP sites characteristic of the suspected pollution source.

10. The method according to claim 9, characterized in that, The criteria for determining the true source of pollution based on similarity are as follows: when the similarity of the samples to be identified exceeds 60%, 70%, 80%, 90%, or 100%, the pollution source is considered to be a true source of pollution. If the sample size is ≥ the third preset threshold, then batch contamination is considered to have occurred, and the suspected contamination source sample is the contamination source. Preferably, the third preset threshold is 0.6, 0.7, 0.8, 0.9, or 1.

0.

11. The method according to claim 7, characterized in that, When the contamination is inter-sample contamination, the method further includes: A3 filters out contaminated sequences to determine whether they represent a true infection. Preferably, in step A3, filtering the contaminating sequence includes: D1, for each contaminated sample other than the contamination source, a sequence carrying ≥2 characteristic iSNPs of the identified contamination source on one of its sequencing sequences is marked as a contaminated sequence and filtered to obtain the filtered sequence; and D2, when the number of filtered sequences is ≥10, the sample to be identified is a true infection.

12. An apparatus for identifying sample contamination in high-throughput sequencing, characterized in that, The device includes: Sequencing information acquisition module: configured to acquire sequencing information obtained after high-throughput sequencing of the sample to be tested; Sample status determination module: configured to determine the sample contamination status according to the method described in any one of claims 7-11 using the characteristic iSNP database for identifying sample contamination in high-throughput sequencing obtained by the construction method as described in any one of claims 1-6.

Images