Infectious outbreak analysis using long-read sequencing data

Abstract

A system and method for outbreak analysis by performing a taxonomic classification of the first LRS data to assign one or more taxonomic identifiers to each record of the first LRS data, identifying a plurality of species-specific candidate genomes from a reference genome database based on the assigned taxonomic identifiers, aligning the first LRS data with each of the plurality of species-specific candidate genomes to determine a first species identity of a microbial species present in the sample, and identifying a first closest strain of the identified microbial species present in the first sample by aligning the first LRS data with a plurality of strain-specific candidate genome datasets.

Description

[Technical Field]

[0001] The present disclosure generally relates to methods and systems for the analysis of infectious outbreaks using long-read sequencing data. [Background technology]

[0002] This background description is provided to generally present the contents of the present disclosure. Nothing in this background section is admitted expressly or implicitly as prior art to the present disclosure.

[0003] Infectious diseases, including those in clinical settings, can have diverse origins. In addition, microorganisms causing infections can follow different evolutionary trajectories as they spread. Infectious diseases can include healthcare-associated infections (HAIs), which include infections that were not present or harbored in patients at the time of admission but that patients acquire during the course of their treatment in a hospital or other healthcare facility. HAIs are associated with significant morbidity and excess mortality, prolonged hospital stays, long-term disabilities, increased antimicrobial resistance, excess mortality, and high healthcare costs. Given their infectious nature and association with antimicrobial resistance, HAIs are a major public health concern. Effective HAI detection and mitigation can protect patients from HAI-related harm and allow scarce healthcare resources to be redirected to other areas of need.

[0004] Traditional laboratory-based methods for revealing relatedness among bacterial isolates include pulsed-field gel electrophoresis (PFGE). Recently, molecular-based methods such as next-generation sequencing (NGS) have been adopted for outbreak investigations. NGS data enable multilocus sequence typing (MLST) and genome assembly, a method for obtaining very long DNA sequences (supercontigs or chromosomes) for pathogen DNA / RNA comparison. However, traditional methods require specialized equipment and expertise. Furthermore, traditional methods are technically cumbersome, requiring different restriction enzymes (PFGE), different sequencing primers for each different species involved (MLST), or high computational resources (genome assembly).

[0005] Although outbreaks are generally thought to be caused by the spread of bacterial strains, the rapid spread of AMR strains can be caused by horizontal gene transfer. No efficient method exists for rapidly analyzing potential plasmid outbreaks, in which clinically important AMR genes spread across species via horizontal gene transfer. With AMR increasing worldwide, rapid analysis of AMR genes carried by both chromosomes and plasmids is essential for optimizing antimicrobial use. When NGS is used to investigate outbreaks, human bioinformatics expertise is required for sequence analysis. Due to these limitations, samples from suspected outbreaks are often submitted to reference laboratories for specialized analysis, resulting in delayed mitigation measures. Summary of the Invention [Problem to be solved by the invention]

[0006] It is desired to address or ameliorate one or more of the disadvantages or limitations associated with the prior art, or at least provide a useful alternative. [Means for solving the problem]

[0007] Some embodiments provide a system for outbreak analysis, comprising: receiving long read nucleic acid sequence data (first LRS data) from a first sample; performing a taxonomic classification of the first LRS data to assign one or more taxonomic identifiers to each record of the first LRS data; Identifying a plurality of species-specific candidate genomes from a reference genome database based on the assigned taxonomic identifiers; aligning the first LRS data with each of a plurality of species-specific candidate genomes to determine a first species identity of the microbial species present in the sample; Identifying a first closest strain of the identified microbial species present in the first sample by aligning the first LRS data with the multiple strain-specific candidate genome datasets. The present invention relates to a system including one or more processing units for:

[0008] Identifying the first closest strain may include aligning a first subset of LRS data corresponding to the first species with a plurality of candidate strain genomes of the first species, calculating a genomic overlap associated with the first subset of LRS data for each of the plurality of candidate strain genomes, and identifying the candidate strain as the first closest strain in the first sample based on the highest genomic overlap value.

[0009] In some embodiments, the one or more processing units are further configured to monitor the abundance of the taxonomic identifiers assigned to the first LRS data, and in response to the abundance of the subset of the taxonomic identifiers being stable, select a subset of species-specific candidate genomes from the reference genome database.

[0010] In some embodiments, the one or more processing units are further configured to receive long read nucleic acid fragment sequence data (second LRS data) from a second sample, perform taxonomic classification of the second LRS data to assign one or more taxonomic identifiers to each record of the second LRS data, identify a subset of the second LRS data that corresponds to a first species identity based on the assigned taxonomic classifiers, and align the second LRS data with multiple strain-specific candidate genome datasets to identify a second closest strain in the second sample.

[0011] In some embodiments, the one or more processing units are further configured to determine a similarity or difference between the first closest strain and the second closest strain to identify an infection outbreak.

[0012] In some embodiments, the one or more processing units are further configured to input a phylogenetic structure based on the determined similarity or difference between the first closest strain and the second closest strain.

[0013] In some embodiments, the one or more processing units are further configured to align records of the first LRS data to determine a first contig, where the contig represents a chromosome or a plasmid of a microorganism in the first sample, align the first contig with the first draft genome by correcting alignment errors of the records of the first LRS data to construct a first consensus genome, and detect a sequence of interest in the first consensus genome by comparing the first consensus genome with a database of known sequences of interest.

[0014] In some embodiments, the sequence of interest comprises an antimicrobial resistance (AMR) gene or a virulence factor sequence.

[0015] Comparing the consensus genome to a database of known sequences of interest can include determining an identity score, or sequence coverage score, for each record in the database of known sequences of interest with the consensus genome.

[0016] In some embodiments, the one or more processing units are further configured to receive long read nucleic acid fragment sequence data (second LRS data) from a second sample, align records of the second LRS data to determine second contigs, where the contigs represent chromosomes or plasmids of microorganisms in the second sample, align the second contigs with the second draft genome by correcting alignment errors of the records of the second LRS data to construct a second consensus genome, and detect a sequence of interest in the second consensus genome by comparing the second consensus genome with a database of known sequences of interest.

[0017] In some embodiments, the one or more processing units are further configured to compare the sequence of interest detected in the first consensus genome with the sequence of interest detected in the second consensus genome to detect horizontal transfer of any one of the sequences of interest between the microorganisms in the first and second samples.

[0018] Some embodiments relate to a method for infection outbreak analysis, comprising determining long-read nucleic acid sequence data (first LRS data) from a first sample; performing a taxonomic classification of the first LRS data to assign one or more taxonomic identifiers to each record of the first LRS data; identifying a plurality of species-specific candidate genomes from a reference genome database based on the assigned taxonomic identifiers; aligning the first LRS data with each of the plurality of species-specific candidate genomes to determine a first species identity of a microbial species present in the sample; and identifying a first closest strain of the identified microbial species present in the first sample by aligning the first LRS data with a plurality of strain-specific candidate genome datasets.

[0019] The methods of some embodiments further comprise subjecting the first sample to selective culture growth conditions to select for a particular group of microorganisms for infection outbreak analysis.

[0020] The method of some embodiments further comprises monitoring the abundance of taxonomic identifiers assigned to the first LRS data, and selecting a subset of species-specific candidate genomes in response to the abundance of the subset of taxonomic identifiers being stable.

[0021] The method of some embodiments further includes determining long read nucleic acid fragment sequence data (second LRS data) from the second sample; performing taxonomic classification of the second LRS data to assign one or more taxonomic identifiers to each record of the second LRS data; identifying a subset of the second LRS data that corresponds to a first species identity based on the assigned taxonomic classifiers; and aligning the second LRS data with multiple strain-specific candidate genome datasets to identify a second closest strain in the second sample.

[0022] The methods of some embodiments further comprise determining a similarity or difference between the first closest strain and the second closest strain to identify the infection outbreak.

[0023] The method of some embodiments further comprises inputting a phylogenetic structure based on the determined similarity or difference between the first closest strain and the second closest strain.

[0024] In some embodiments, the method further includes aligning records of the first LRS data to determine a first contig, where the contig represents a chromosome or a plasmid of a microorganism in the first sample; aligning the first contig with the first draft genome by correcting alignment errors of the records of the first LRS data to construct a first consensus genome; and comparing the first consensus genome with a database of known sequences of interest to detect the sequence of interest in the first consensus genome.

[0025] The sequence of interest may include an antimicrobial resistance (AMR) gene or a virulence factor sequence.

[0026] Comparing the consensus genome to a database of known sequences of interest can include using the consensus genome to determine an identity score or sequence coverage score for each record in the database of known sequences of interest.

[0027] In some embodiments, the method further includes determining long read nucleic acid fragment sequence data (second LRS data) from a second sample; aligning records of the second LRS data to determine second contigs, where the second contigs represent chromosomes or plasmids of microorganisms in the second sample; aligning the second contigs with the second draft genome by correcting alignment errors of records of the first LRS data to construct a second consensus genome; and comparing the second consensus genome with a database of known sequences of interest to detect the sequence of interest in the second consensus genome.

[0028] The methods of some embodiments further include comparing the sequence of interest detected in the first consensus genome with the sequence of interest detected in the second consensus genome to detect horizontal transfer of any one of the sequences of interest between the microorganisms in the first and second samples.

[0029] Exemplary embodiments of the present invention are illustrated by way of example in the accompanying drawings, in which like reference numerals indicate the same or similar elements, and in which: [Brief explanation of the drawings]

[0030] [Figure 1] 1 is a schematic diagram illustrating a portion of a method according to the present disclosure. [Figure 2] FIG. 2 is another schematic diagram illustrating a portion of a method according to the present disclosure. [Figure 3] FIG. 1 is a schematic diagram showing a comparison of a method according to the present disclosure with a baseline method for outbreak tracking. [Figure 4] FIG. 10 illustrates a cluster map showing some of the outbreak tracking results obtained by embodiments according to the present disclosure. [Figure 5] FIG. 1 is a block diagram of a system for outbreak analysis. DETAILED DESCRIPTION OF THE INVENTION

[0031] Embodiments provide systems and methods for rapidly detecting or tracking potential outbreaks using sequence-based techniques that can determine the genetic relatedness of chromosomes or plasmids. Some embodiments also identify AMR genes and other potential virulence genes carried by chromosomes or plasmids in a sample. Embodiments include a computer system or platform for providing analytical services on long-read sequencing data in the context of outbreak analysis. The system or platform can provide output in the form of analytical reports, including phylogenetic trees, strain / plasmid similarity maps, and strain-specific cluster reports, to assist in outbreak analysis at facilities, such as healthcare facilities.

[0032] Rapid detection and mitigation of infectious diseases, including HAIs, requires both epidemiological and laboratory data. When an outbreak is suspected, laboratories perform cultures (from relevant clinical and environmental samples) and obtain bacterial isolates for downstream analysis to determine relatedness. Assessing the relationships between outbreak isolates is important to understanding the outbreak. Relationships between outbreak isolates may enable the identification of previously unknown routes of transmission of infectious diseases in healthcare settings. With knowledge of the routes of transmission, embodiments generate instructions for taking preventative measures to control the spread of the infection.

[0033] Embodiments enable outbreak analysis and / or tracking by providing computational models and methods for processing long-read sequencing information (LRS data) obtained from nanopore DNA sequencing platforms or other similar platforms that provide long-read nucleic acid data (LRS data). Algorithms of embodiments can distinguish strains with an average nucleotide difference of 0.01% or less, facilitating the identification and / or differentiation of outbreak clusters. Some embodiments also identify antimicrobial resistance (AMR) genes in bacterial plasmids to gain insight into the horizontal transmission of antimicrobial resistance traits across bacterial isolates. Embodiments provide DNA sequencing protocols and computational models for processing long-read data to provide insightful information to facilitate analysis of outbreaks in clinical settings, such as hospitals, as shown in FIG. 1 .

[0034] Embodiments can be integrated with laboratory or clinical protocols, and embodiment algorithms utilize long-read sequencing data to provide insightful information for outbreak analysis, as shown in FIG. 2. Some embodiments are described with reference to a single sample or two samples. The methods and computational steps of the embodiments can be extended to process multiple samples obtained from multiple locations in a clinical setting. Outbreak analysis results, such as strain identification, strain-specific clustering, comparison between two strains, analysis of strain / plasmid similarity, and generation of similarity maps / trees, can be extended to multiple samples. Reference to a single sample or first / second sample is merely for convenience and is not intended to limit the scope of the claimed embodiments.

[0035] Obtaining pathogen isolates from clinical samples Routine clinical samples, such as bronchoalveolar lavage (BAL), screening swabs, and blood culture samples, are collected from patients according to routine clinical practice (step 210 in Figure 2). Samples are collected aseptically in sterile containers and transported to a diagnostic laboratory within the hospital for processing. In some cases, clinical samples can be processed within an hour of collection. Pathogens are isolated from the sample by a series of selective culture growth conditions to select for specific bacterial groups.

[0036] DNA extraction and nanopore library preparation The disclosed embodiments are applicable to any DNA sequencing platform capable of generating sequencing data containing "reads" longer than 1,000 bases in length. Some embodiments use nanopore sequencing technology to generate sequencing data (first / second LRS data). Total nucleic acid extraction (step 214 in FIG. 2 ) and library preparation (step 216 in FIG. 2 ) can be performed according to nanopore long-read sequencing protocols (e.g., SQK-LSK109 / LSK110). When a MinION flow cell is used, up to 24 or 96 samples (depending on the barcoding kit selected) can be sequenced in a single run to generate LRS data.

[0037] Computational Module for Pathogen Outbreak Tracking Figure 5 illustrates an outbreak analysis system 500 and its associated components. Once sequencing of a sample 505 begins (step 220 of Figure 2), raw sequencing signals are converted locally to DNA sequences (reads) in real time. Demultiplexed reads (LRS data 545) are streamed to the outbreak analysis system 500. The present disclosure contemplates any suitable number of systems 500. The present disclosure contemplates the computer system 500 taking any suitable physical form. By way of example, and not limitation, the computer system 500 may be an embedded computer system, a system-on-chip (SOC), a single-board computer system (SBC) (e.g., a computer-on-module (COM) or system-on-module (SOM)), a desktop computer system, a laptop or notebook computer system, a mainframe, a mesh of computer systems, a mobile phone, a server, a tablet computer system, or a combination of two or more thereof. Where appropriate, the computer system 500 can include one or more computer systems 500, and can be single or distributed, spanning multiple locations, spanning multiple machines, spanning multiple data centers, or residing in a cloud that can include one or more cloud components in one or more networks. Where appropriate, the one or more computer systems 500 can perform one or more steps of one or more methods described or exemplified herein without substantial spatial or temporal limitations. Where appropriate, the one or more computer systems 500 can perform one or more steps of one or more methods described or exemplified herein at different times or in different locations. The system 500 enables automated and scalable analysis of various numbers of samples at different times. Two analysis modules of an embodiment are described below.

[0038] Rapid closest strain identification module The memory 520 of the system 500 includes program code for processing the LRS data 545 generated by the sequencing platform 540. The streaming-in reads (raw sequencing signals, LRS data 545) are classified into specific species using a taxonomic classifier 522 (step 222 of FIG. 2 ) based on a curated pan-genome database (reference genome database 550). Depending on the sequencing throughput, when the abundance of a species in the raw sequencing signal from a sample reaches a certain threshold (e.g., the total number of nucleotides identified in the sample is equal to or greater than the genome size of a given species), embodiments can retrieve or query species-specific records from the reference genome database for subsequent analysis. Records of the LRS data are classified based on the pan-genome database and / or the species-specific database, or both databases simultaneously. The taxonomic classifier monitors species-specific abundance. If abundance is stable, 10-20 candidate genomes are retrieved from the reference genome database for long-read alignment. Abundance monitoring is performed by system 500 based on the output of the taxonomic classifier once LRS data is received by system 500. Abundance monitoring over an extended analysis period helps avoid abundance biases that may occur at the beginning of the sequencing process, resulting in more accurate strain identification.

[0039] Processing of the LRS data can be performed over several intervals. For each interval, embodiments obtain LRS data that can be stored in a FASTQ-formatted file, including multiple "reads," i.e., DNA fragments of different lengths (1,000 to 100,000 nucleotides). A K-mer profile is extracted for each record of the LRS data. A K-mer refers to all subsequences of a read with length K, where K ranges from 3 to 31 nucleotides.

[0040] The taxonomic classifier 522 assigns one or more taxonomic identifiers to each read based on the K-mer profile and comparison with records in a reference genome database 550 accessible to the taxonomic classifier 522. The taxonomic identifiers represent operational taxonomic units (OTUs), which may refer to domains, kingdoms, phyla, classes, orders, families, genus, species, strains, or individual genomes. The reference genome database contains DNA sequences of microbial genomes of interest for outbreak tracking. The taxonomic classifier incrementally updates counts for each OTU (species level) as LRS data is received by the system 500. The taxonomic classifier monitors OTU counts and triggers coverage analysis by the coverage analyzer 526 for each species when the species count exceeds a threshold. Species with counts higher than the threshold in the replicates may be referred to as candidate species. The threshold may be defined based on the total number of sequenced nucleotides and the genome size of each species.

[0041] The alignment module 524 performs an alignment of the LRS data associated with the candidate species with the associated species-specific reference genome. Based on the long read alignment results, a coverage analysis is performed by the coverage analyzer 526 to confirm the presence of one or more species in the sample 505.

[0042] The system 500 selects candidate reference genomes for each species based on OTU counts at the strain or individual genome level. The alignment module aligns the classified LRS data to the candidate reference genome and records the genomic regions covered by the reads. A read can be said to align with a genome if the identity score (total number of matched nucleotides / alignment length × 100) is at least 80%, 85%, or 90% and / or the read coverage score (alignment length / read length × 100) is at least 80%, 85%, or 90%. Given the alignment records, the coverage analyzer calculates the coverage breadth percentage for each candidate species. The coverage analyzer can confirm the presence of a species if the coverage percentage is at least 40-90% of the expected coverage for each species. The expected coverage percentage can follow a Poisson distribution that takes into account the total number of sequenced nucleotides and the genome size of each species. In alternative embodiments, other suitable distributions may be used by the coverage analyzer, including the negative binomial distribution.

[0043] Strain-level taxonomic classification Long read alignment also allows for the identification of unique regions in the LRS data that provide the basis for determining the closest strain to the microorganism identified in the sample (step 226 in Figure 2). As the LRS data is received by system 500, system 500 updates species and / or strain identification results in real time or near real time. The generated results may be displayed on display 560.

[0044] For each detected species, the entity searches the reference genome and calculates the genome similarity based on the whole genome information. The reference genome is then classified based on the similarity value. In some embodiments, the similarity value may be 99%, 99.5%, 99.9%, 99.95%, or 99.99%. For each cluster / group defined based on the similarity value, a representative genome is selected from the reference genome. A species-specific dataset is defined for taxonomic classification based on the selected representative genome. Taxonomic classification related to the species-specific dataset is performed as described above, except that the OTUs used for taxonomic classification are strain-level OTUs or related to specific known strains of microorganisms. The number of candidate genomes in the strain-specific dataset may be 10, 15, or 20 or less. Reducing the search space of candidate genomes in the strain-specific dataset based on previous coverage analysis using pan-genome records in the reference genome database advantageously reduces the computational complexity of the strain identification operation. Based on the alignment results from the coverage analysis operation, embodiments identify reads that are uniquely aligned to a single candidate reference genome in the strain-specific dataset.

[0045] For each candidate representative genome in the strain-specific dataset, the embodiment determines the percentage of the genome region covered by uniquely aligned reads (also referred to as genome overlap).The system 500 then compares the genome overlap values ​​from all candidate representative genomes and identifies the closest strain, i.e., the candidate representative genome with the highest overlap percentage.If the highest percentage is significantly higher than the average overlap percentage across all candidate representative genomes, the closest strain is said to be significantly matched.

[0046] Detection and identification of AMR genes The system 500 may also include an AMR gene analysis module 529. The AMR gene analysis module performs genome assembly (step 228 in Figure 2) and polishing to obtain a high-quality draft genome consisting of full-length (or near-full-length) chromosome and plasmid sequences. The assembly process may begin when all samples reach a certain sequencing depth (>50-100x) or after sequencing is complete. All assembled sequences are then aligned using a pairwise genome alignment method. The reads are aligned to an AMR gene database that is part of the draft genome and / or reference genome database 550. The AMR analysis module then identifies AMR genes in the LRS data and reports the chromosomes or plasmids that carry the AMR genes, as shown in the example in Table 1.

[0047] By integrating the pairwise alignment of plasmids and the alignment results of AMR genes, embodiments allow pinpointing of plasmids containing AMR genes of interest (e.g., input by a user of system 500; such as the OXA family). A plasmid similarity tree containing AMR is generated by system 500, providing insight into horizontal plasmid transfer within an outbreak, as shown in FIG. 4.

[0048] Long reads obtained from a sample (e.g., received in fastq format) are aligned to each other, and the concatenation of all reads (i.e., contigs) is reported or identified by the system 500. Each contig may represent a chromosome or a plasmid. If a plasmid is present in the isolate, or if the genome assembly process was unable to align all chromosomal reads into a single contig, there may be more than one contig associated with the sample. The single contig or set of contigs from each isolate sample is referred to as the draft genome.

[0049] Embodiments align a subset or all of the long reads obtained from a sample to a draft genome to build a consensus genome. Any errors in the LRS data are corrected during the assembly process. The polishing process (alignment of all long reads) can be performed multiple times to reduce errors. Embodiments determine the resulting consensus genome. Contig polishing involves mapping the LRS data from the sample to contigs and building a consensus of this read mapping to resolve putative assembly errors in the contigs.

[0050] The system 500 identifies whether each contig is chromosomal or plasmid-derived by examining the mobile / plasmid elements within the contig. Based on chromosomal contigs from multiple isolates / samples, the system 500 can perform a whole-chromosome comparison or phylogenetic analysis. Based on contigs derived from plasmids from multiple isolates / samples, a whole-plasmid comparison or phylogenetic analysis can also be performed. Phylogenetic analysis involves generating a phylogenetic structure / tree (112 in Figures 1, 2, and 3) that shows the evolutionary relationships among multiple strains identified from multiple samples. Phylogenetic analysis can also include analyzing the similarity between AMR-containing plasmids sequenced and identified from multiple samples. The AMR similarity analysis can be represented in the form of a similarity heat map (114 in Figures 1 and 2).

[0051] Embodiments detect AMR genes and pathogenic factors (i.e., sequences of interest) in each polished contig by aligning the sequences of the AMR and pathogenic factors on the contig. Open reading frames may also be identified as part of the sequence of interest detection process. For example, a sequence of interest may be considered to align with an AMR gene if the identity score is at least 90%, 95%, or 99%, and / or the sequence coverage score (alignment length / sequence length x 100) is at least 80%, 85%, or 90%.

[0052] Based on the AMR genes detected on the plasmid, horizontal gene transfer can be detected in the event that the same AMR gene is detected in different isolates. Long reads can also be directly aligned to the sequence of interest by the system 500. A read is said to align with an AMR gene if the identity score is at least 80%, 85%, or 90% and the sequence coverage score (alignment length / gene length × 100) is at least 80%, 85%, or 90%. Embodiments can report a list of AMR genes, virulence factors, and horizontal gene transfer events detected within the input isolates. Horizontal gene transfer (HGT) involves the asexual transfer of genetic information between the genomes of microorganisms in two samples. HGT involves the spread of antibiotic resistance genes between bacteria, and detecting HGT between samples can provide insight into infectious disease outbreaks within a facility. Detecting HGT allows for evaluation of the effectiveness of treatment modalities and, if necessary, modification of treatment modalities or implementation of infection control measures.

[0053] Comparison of different techniques for outbreak tracking Results obtained using embodiments were compared with existing outbreak analysis techniques using 52 Klebsiella pneumoniae isolates obtained between 2018 and 2020 from both clinical and environmental samples. Figure 2 shows a schematic diagram of the flow chart for closest match strain identification and outbreak tracking using embodiments. Existing outbreak tracking techniques, including pulsed-field gel electrophoresis (PFGE) and next-generation sequencing (NGS) based on the Illumina platform, were used as baselines for comparison. NGS data was analyzed using two baseline methods.

[0054] NGS-MLST: Multilocus sequence typing (MLST) compares Illumina short reads to a standard typing scheme provided by PubMLST (Jolley, Bray & Maiden, 2018), then identifies sequence types (STs) using SRST2 (Inouye et al., 2014).

[0055] NGS-ASM: Assemble draft genomes using Illumina short reads (GERMS pipeline in GIS). Create phylogenetic trees based on the distances between core genomes of multiple isolates (Treangen, Ondov, Koren & Phillippy, 2014).

[0056] All baseline methods suggest distinct outbreak clusters across different years. NGS-MLST and NGS-ASM provide similar outbreak cluster information, but NGS-MLST failed to specifically cluster some samples. The closest strains identified based on the disclosed embodiments were consistent with the PFGE-based results.

[0057] Figure 3 shows a comparison between different outbreak tracking techniques. The phylogenetic tree in Figure 3 was constructed based on NGS-ASM. Each node represents a Klebsiella pneumoniae isolate, and the labels indicate the collection year, location, sample site, PFGE type, MLST (NF = not found; * = mismatch; ? = uncertain; nan = no core gene alignment), and the ID of the closest strain identified by the disclosed embodiments (9999XXX). As shown in Figure 3, the strain classification generated by the embodiments (the 9999XXX associated with each isolate) is consistent with the phylogenetic tree generated based on NGS-ASM.

[0058] In addition to sequencing analysis based on core genome / chromosomal information, some embodiments offer the advantage of providing the ability to assess whether AMR genes are spreading within an outbreak. Experiments have observed that the majority of AMR gene families reside on plasmids and are transferable to other bacteria. While PFGE and NGS-based methods cannot detect transferable plasmids, the AMR analysis module of system 500 can identify and characterize AMR genes in samples. Similarities between plasmids across different bacterial isolates were calculated, providing unique high-resolution information for outbreak tracking, as illustrated by the results shown in Figure 4. Experiments have detected and identified OXA-48-like genes on plasmids in 51 isolates (out of 52 isolates in which OXA-48-like genes were detected by qPCR). As expected, this experiment detected unique plasmid types across different years in most cases.

[0059] Figure 4 shows a cluster map illustrating the similarity between plasmids containing AMR genes across 51 Klebsiella pneumoniae isolates. Each row label includes the sample ID, contig ID, contig / plasmid size, and the detected OXA gene. Shading represents the plasmid similarity. The system 500 can generate such a similarity heat map and display the generated heat map on the display 560.

[0060] Integration into routine outbreak tracking in hospitals The combined results of some embodiments are comprised of outbreak insights based on real-time analysis of nanopore long-read sequencing data in Module 1 and Module 2, as shown in Table 1 below, highlighting the different outbreak perspectives provided by the results of the embodiments.

[0061] [Table 1A] [Table 1B] [Table 1C] [Table 1D] [Table 1E]

[0062] Some embodiments rely on both cloud computing / remote computing platforms (for highly scalable, real-time whole-genome analysis) and local in-lab computers for live base calling (converting raw sequencing signals into DNA sequences). In some embodiment setups, when 24 isolates are sequenced on a MinION flow cell, experiments can identify the closest match (Rapid Tax ID column) for each isolate within 1-2 hours and characterize the OXA-containing plasmid within a day. Exemplary hardware and software specifications, as well as reagents, are listed in Table 2.

[0063] [Table 2A] [Table 2B]

[0064] It will be appreciated that many further modifications and permutations of various aspects of the described embodiments are possible, and accordingly, the described aspects are intended to embrace all such alterations, modifications, and variations that fall within the spirit and scope of the appended claims.

[0065] Long-read nucleic acid fragment sequence (LRS) data includes sequencing data of at least 1,000 base pairs or more of a DNA or RNA molecule. Long-read nucleic acid fragment sequence data can be obtained using nanopore sequencing, or PacBio sequencing, or any other long-read sequencing technology.

[0066] The predefined abundance level includes a level of abundance that is considered statistically significant in terms of identifying a microorganism in a sample. The predefined abundance level may include a level where the total number of sequenced bases in the sample is equal to or greater than the genome size of a given species.

[0067] The reference genomes include genomes corresponding to various species that may be present in the sample. The reference genomes may be stored in a genome database populated by routine clinical analysis of samples by total nucleic acid extraction and long-lead ligation. The candidate genomes include a subset of the reference genomes selected for coverage analysis. The candidate genomes are selected based on taxonomic identifiers assigned to the LRS data obtained from the sample. The selection of the candidate genomes advantageously avoids the need to align large amounts of LRS data with multiple reference genomes, which would increase the computational cost of the method of the embodiment. The candidate genomes may also be referred to as representative genomes or candidate reference genomes.

[0068] Aligning the LRS data with the candidate genome involves matching nucleotides of the LRS data with the genome.

number

number

[0069] Coverage analysis involves calculating the percentage of coverage breadth for each candidate genome based on the alignment results.

[0070] Species identity includes sufficient genomic data or portions of genomic data to uniquely identify different species of microorganisms in a sample.

[0071] Some embodiments are directed to a method / system for outbreak analysis by determining long read nucleic acid fragment sequence data (LRS data) from a plurality of samples, performing a taxonomic classification of the LRS data and comparison with a plurality of candidate genomes to determine a plurality of species present in the plurality of samples, aligning the LRS data of the plurality of species with a candidate strain genome database to determine the closest strain of each of the plurality of species, and determining the similarity or difference between the determined closest strains to identify infection outbreaks.

[0072] Some embodiments are directed to a system / method for determining a consensus genome by determining long read nucleic acid fragment sequence data (LRS data) from a sample; aligning records of the LRS data to determine one or more contigs, where the contigs represent chromosomes or plasmids; and aligning the contigs with a draft genome by correcting alignment errors in the records of the LRS data to construct a consensus genome of the microorganisms of the sample.

[0073] Some embodiments relate to a system / method for detecting a sequence of interest in a microorganism, the method comprising: determining long read nucleic acid fragment sequence data (first LRS data) from a first sample; aligning records of the first LRS data to determine at least one first contig, where the contig represents a chromosome or a plasmid of the microorganism in the first sample; aligning the at least one first contig with a first draft genome by correcting alignment errors of the records of the first LRS data to construct a first consensus genome; and comparing the first consensus genome with a database of known sequences of interest to detect the sequence of interest in the first consensus genome.

[0074] Some embodiments relate to a method for treating an infection caused by one or more microorganisms in a subject, the method comprising determining long read nucleic acid fragment sequence data (LRS data) from a sample; performing a taxonomic classification of the LRS data to assign one or more taxonomic identifiers to each record in the LRS data; identifying a plurality of candidate genomes based on the taxonomic identifiers assigned to each record in the LRS data; aligning the LRS data with each of the plurality of candidate genomes to determine the species identity of the microorganism present in the sample; identifying a closest strain to the microorganism based on the aligned LRS data and a candidate strain genome database; and administering a therapeutic agent to the subject based on the identified closest strain.

[0075] Some embodiments relate to a system / method for outbreak tracking by determining long read nucleic acid fragment sequence data (LRS data) from a plurality of samples, performing a taxonomic classification of the LRS data and comparison with a plurality of candidate genomes to determine a plurality of species present in the plurality of samples, aligning the LRS data of the plurality of species with a candidate strain genome database to determine the closest strain of each of the plurality of species, determining the similarity or difference between the determined closest strains to identify an infection outbreak, aligning a first LRS data record to determine at least one contig, where the contig represents a chromosome or a plasmid of a microorganism in one of the plurality of samples, aligning the at least one contig with a draft genome by correcting alignment errors of the LRS data records to construct a consensus genome, and comparing the consensus genome with a database of known sequences of interest to detect sequences of interest in the consensus genome.

[0076] Throughout this specification and the claims that follow, unless the context requires otherwise, the word "comprise" and variations such as "comprises" and "comprising" will be understood to mean the inclusion of a stated integer or step or group of integers or steps, but not to the exclusion of any other integer or step or group of integers or steps.

[0077] Any reference herein to any prior publication (or information derived therefrom), or to any publicly known matter, is not, and should not be taken as, a form of acknowledgment or admission or any suggestion that the prior publication (or information derived therefrom) or publicly known matter forms part of the common general knowledge in the field of endeavor to which this specification pertains.

Claims

1. 1. A system for outbreak analysis, comprising: receiving long read nucleic acid sequence data (first LRS data) from a first sample; performing a taxonomic classification of the first LRS data to assign one or more taxonomic identifiers to each record of the first LRS data; Identifying a plurality of species-specific candidate genomes from a reference genome database based on the assigned taxonomic identifiers; aligning the first LRS data with each of the plurality of species-specific candidate genomes to determine a first species identity of a microbial species present in the sample; Identifying a first closest strain of the identified microbial species present in the first sample by aligning the first LRS data with a plurality of strain-specific candidate genome datasets. The system includes one or more processing units for:

2. Identifying the first closest strain aligning the subset of the first LRS data corresponding to the first species with a plurality of candidate strain genomes of the first species; calculating, for each of the plurality of candidate strain genomes, a genomic overlap associated with the subset of the first LRS data; identifying a candidate strain as the first closest strain in the first sample based on the highest genome overlap value; The system of claim 1 , comprising:

3. the one or more processing units: monitoring the abundance of taxonomic identifiers assigned to the first LRS data; Selecting a subset of species-specific candidate genomes from a reference genome database in response to the abundance of the subset of taxonomic identifiers being stable. The system of claim 1 , further configured to:

4. the one or more processing units: receiving long read nucleic acid fragment sequence data (second LRS data) from a second sample; performing a taxonomic classification of the second LRS data to assign one or more taxonomic identifiers to each record of the second LRS data; identifying a subset of the second LRS data that corresponds to a first species identity based on an assigned taxonomic classifier; Aligning the second LRS data with the plurality of strain-specific candidate genome datasets to identify a second closest strain in the second sample. The system of claim 1 , further configured to:

5. 5. The system of claim 4, wherein the one or more processing units are further configured to determine a similarity or difference between the first closest strain and a second closest strain to identify an infection outbreak.

6. 5. The system of claim 4, wherein the one or more processing units are further configured to input a phylogenetic structure based on the determined similarity or difference between the first closest strain and the second closest strain.

7. the one or more processing units: aligning records of the first LRS data to determine a first contig, the contig representing a chromosome or a plasmid of a microorganism in the first sample; aligning the first contig with a first draft genome by correcting alignment errors of records of the first LRS data to construct a first consensus genome; Detecting a sequence of interest in the first consensus genome by comparing the first consensus genome to a database of known sequences of interest. The system of claim 1 , further configured to:

8. The system of claim 7, wherein the sequence of interest comprises an antimicrobial resistance (AMR) gene or a virulence factor sequence.

9. 8. The system of claim 7, wherein comparing the consensus genome to the database of known sequences of interest comprises determining an identity score or sequence coverage score for each record in the database of known sequences of interest with the consensus genome.

10. the one or more processing units: receiving long read nucleic acid fragment sequence data (second LRS data) from a second sample; aligning records of the second LRS data to determine second contigs, the contigs representing chromosomes or plasmids of microorganisms in the second sample; aligning the second contig with a second draft genome by correcting alignment errors of records of the second LRS data to construct a second consensus genome; Detecting a sequence of interest in the second consensus genome by comparing the second consensus genome to a database of known sequences of interest. The system of claim 7, further configured to:

11. 11. The system of claim 10, wherein the one or more processing units are further configured to compare the sequence of interest detected in the first consensus genome with the sequence of interest detected in the second consensus genome to detect horizontal transfer of any one of the sequences of interest between microorganisms in the first and second samples.

12. 1. A method for infection outbreak analysis, comprising: determining long read nucleic acid sequence data (first LRS data) from a first sample; performing a taxonomic classification of the first LRS data to assign one or more taxonomic identifiers to each record of the first LRS data; identifying a plurality of species-specific candidate genomes from a reference genome database based on the assigned taxonomic identifiers; aligning the first LRS data with each of the plurality of species-specific candidate genomes to determine a first species identity of a microbial species present in the sample; identifying a first closest strain of the identified microbial species present in the first sample by aligning the first LRS data with a plurality of strain-specific candidate genome datasets. A method comprising:

13. identifying a first closest strain, aligning the first subset of LRS data corresponding to the first species with a plurality of candidate strain genomes of the first species; calculating, for each of the plurality of candidate strain genomes, a genomic overlap associated with the subset of the first LRS data; identifying a candidate strain as the first closest strain in the first sample based on the highest genome overlap value; 13. The method of claim 12, comprising:

14. 13. The method of claim 12, further comprising subjecting the first sample to selective culture growth conditions to select for a particular group of microorganisms for infection outbreak analysis.

15. monitoring the abundance of taxonomic identifiers assigned to said first LRS data; Selecting a subset of species-specific candidate genomes in response to the abundance of the subset of taxonomic identifiers being stable. The method of claim 12, further comprising:

16. determining long read nucleic acid fragment sequence data (second LRS data) from a second sample; performing a taxonomic classification of the second LRS data to assign one or more taxonomic identifiers to each record of the second LRS data; identifying a subset of the second LRS data that corresponds to the first species identity based on the assigned taxonomic classifier; aligning the second LRS data with the plurality of strain-specific candidate genome datasets to identify a second closest strain in the second sample; The method of claim 12, further comprising:

17. 17. The method of claim 16, further comprising determining a similarity or difference between the first closest strain and the second closest strain to identify an infection outbreak.

18. 17. The method of claim 16, further comprising the step of inputting a phylogenetic structure based on the determined similarities or differences between the first closest strain and the second closest strain.

19. aligning records of the first LRS data to determine a first contig, the contig representing a chromosome or a plasmid of a microorganism in the first sample; aligning the first contig with a first draft genome by correcting alignment errors of records of the first LRS data to construct a first consensus genome; comparing said first consensus genome to a database of known sequences of interest to detect sequences of interest in said first consensus genome. The method of claim 12, further comprising:

20. 20. The method of claim 19, wherein the sequence of interest comprises an antimicrobial resistance (AMR) gene or a virulence factor sequence.

21. 20. The method of claim 19, wherein comparing the consensus genome to the database of known sequences of interest comprises determining an identity score or sequence coverage score for each record in the database of known sequences of interest with the consensus genome.

22. determining long read nucleic acid fragment sequence data (second LRS data) from a second sample; aligning records of the second LRS data to determine a second contig, the second contig representing a chromosome or a plasmid of a microorganism in the second sample; aligning the second contig with a second draft genome by correcting alignment errors of records of the first LRS data to construct a second consensus genome; comparing said second consensus genome to a database of known sequences of interest to detect sequences of interest in said second consensus genome.

20. The method of claim 19, further comprising:

23. 23. The method of claim 22, further comprising comparing the sequence of interest detected in the first consensus genome with the sequence of interest detected in the second consensus genome to detect horizontal transfer of any one of the sequences of interest between microorganisms in the first and second samples.

Images