Digital standard substance for detecting metagenome biological information detection accuracy

Abstract

The invention provides a digital standard substance for detecting the accuracy of a metagenome biological information detection process. 24 virus genomes which are particularly concerned by Chinese pharmacopoeia are selected, 100 samples are constructed in total, each sample contains genome sequences of 24 different viruses, the abundance is randomly generated, and the virus abundance distribution of each sample is different. The digital standard substance provided by the invention can be used for checking the accuracy of a biological information process established based on a metagenome.

Description

Technical Field

[0001] The present invention relates to the field of bioinformatics, and in particular to digital standards for testing the accuracy of metagenomic bioinformatics detection. Background Art

[0002] Metagenomics extracts the DNA of all microorganisms from microecological samples, constructs a metagenomic library, and then uses genomic research strategies to analyze the genetic composition of all microorganisms in the sample and the potential functions of their communities. This omics technology does not rely on the isolation and pure culture technology of single bacteria, which largely solves the problem that most microorganisms are difficult to study because they cannot be isolated and cultured. At the same time, it can also reflect the true situation of the composition of microorganisms in the ecological environment. In the microecological research on human health, the quantification of microorganisms based on metagenomic sequencing data is the basis for studying their community composition, species interactions, and exploring their relationship with the occurrence and development of diseases. With the progress of scientific research, more and more studies have shown that it is increasingly important to accurately annotate the lower-level taxonomic units of specific species, namely strains. If the relationship between bacteria and diseases is simply studied at a higher taxonomic level, it is likely that the categories that are positively correlated, irrelevant, or even negatively correlated with the development of the disease will be added together, which is obviously wrong both in biology and statistics; and the existing research results are also in urgent need of correction by improving the accuracy of microbial quantification, or conducting more in-depth mechanism research.

[0003] At present, in the biological product testing industry, more and more people are using metagenomic technology to detect exogenous viruses. Through high-throughput metagenomic technology, it is expected that all microorganisms in the sample (bacteria, viruses, fungi, etc.) can be detected at one time, but its subsequent verification work needs to be carried out as soon as possible before this detection technology can be popularized.

[0004] The metagenomics technology developed in recent years bypasses traditional microbial isolation and culture methods and directly extracts total DNA from environmental samples. It obtains new functional genes and bioactive substances by constructing and screening metagenomic libraries. Metagenomic libraries include both culturable and unculturable microbial genetic information, thereby increasing the chance of obtaining new bioactive substances. Its quantitative method has a high resolution, that is, it can be annotated to the species or strain level, so it is currently widely used in pathogen detection.

[0005] In the field of biological product testing, the industry currently mainly uses traditional methods for testing, such as PCR, electrophoresis, and culture methods. However, these traditional methods have low throughput and cannot simultaneously detect a large number of different types of viruses. Therefore, methods based on the second-generation sequencing (NGS) technology route are gradually developing.

[0006] Although the second-generation sequencing can make up for the disadvantage of low throughput, its high sensitivity, high background noise, and randomness of sequencing make it difficult to judge the true and false positives of the test results. At the same time, the sample preparation, DNA and RNA extraction, and subsequent library construction process of the second-generation sequencing are prone to environmental or contamination of consumables. Therefore, while promoting the use of metagenomics for safety testing of biological products, it is necessary to develop standards to test the entire metagenomic workflow, including dry experiments and wet experiments. Among them, the methods and tools of dry experiments (bioinformatics processes) are varied, and there is currently no recognized most reliable detection method and tool for everyone to use. Therefore, it is very necessary to build digital standards to verify the bioinformatics analysis process.

[0007] Therefore, there is a need in the art to develop a digital standard for verifying the accuracy of metagenomic bioinformatics detection. Summary of the invention

[0008] The purpose of the present invention is to provide a digital standard for testing the accuracy of metagenomic biological information detection.

[0009] In a first aspect of the present invention, a method for preparing a digital standard for genomic biological information detection is provided, comprising the following steps:

[0010] (i) providing n reference genomes of microorganisms, wherein n is an integer selected from 10-24, preferably 15-24, more preferably 20-24;

[0011] (ii) simulated community design: generating simulated communities based on the reference genome and taxonomic information;

[0012] (iii) read simulation: using a read simulator to sample from each genome of the simulated community to generate simulated samples, and the simulated samples are mixed to obtain a simulated dataset;

[0013] (iv) Standard assembly: assembling the simulated data set to obtain the digital standard.

[0014] In another preferred embodiment, the microorganism is a virus.

[0015] In another preferred embodiment, the microorganism is n microorganisms selected from the following group: Betapolyomavirusmacacae (simian virus 40), Bluetongue virus (blue tongue virus), Bovine polyomavirus 3 (bovine polyomavirus), Human betaherpesvirus 6A (human herpesvirus-6), Human betaherpesvirus 6B (human herpesvirus-6), Fowl aviadenovirus A (exogenous avian adenovirus type I), Simian retrovirus 8 (monkey retrovirus), Simian foamy virus (monkey foamy virus), Human gammaherpesvirus 4 (human Epstein-Barr virus), Human polyomavirus 12 (human polyomavirus), Human circovirusVS6600022 (human circovirus), Human betaherpesvirus5 (human cytomegalovirus), Humanimmunodeficiency virus 1 (human immunodeficiency virus), Murine respirovirus (Sendai virus), Muridbetaherpesvirus 8 (mouse salivary gland virus_mouse cytomegalovirus), Nipah henipavirus (Nipah virus), Suid alphaherpesvirus 1 (pseudorabies virus), Swinepox virus (swine poxvirus), Porcine reproductive and respiratory syndrome virus (porcine reproductive and respiratory syndrome virus), Suidbetaherpesvirus 2 (porcine cytomegalovirus), Porcine circovirus 2 (porcine circovirus), Humangammaherpesvirus 8 (human herpesvirus-8), Bat rotavirus (mouse rotavirus), Muridbetaherpesvirus 1 (mouse salivary gland virus_mouse cytomegalovirus).

[0016] In another preferred embodiment, the n microorganisms are composed of the following microorganisms: Betapolyomavirusmacacae (simian virus 40), Bluetongue virus (blue tongue virus), Bovine polyomavirus 3 (bovine polyomavirus), Human betaherpesvirus 6A (human herpesvirus-6), Human betaherpesvirus6B (human herpesvirus-6), Fowl aviadenovirus A (exogenous avian adenovirus type I), Simian retrovirus 8 (monkey retrovirus), Simian foamy virus (monkey foamy virus), Human gammaherpesvirus 4 (human Epstein-Barr virus), Human polyomavirus 12 (human polyomavirus), Human circovirusVS6600022 (human circovirus), Human betaherpesvirus5 (human cytomegalovirus), Humanimmunodeficiency virus 1 (human immunodeficiency virus), Murine respirovirus (Sendai virus), Muridbetaherpesvirus 8 (mouse salivary gland virus_mouse cytomegalovirus), Nipah henipavirus (Nipah virus), Suid alphaherpesvirus 1 (pseudorabies virus), Swinepox virus (swine poxvirus), Porcine reproductive and respiratory syndrome virus (porcine reproductive and respiratory syndrome virus), Suidbetaherpesvirus 2 (porcine cytomegalovirus), Porcine circovirus 2 (porcine circovirus), Humangammaherpesvirus 8 (human herpesvirus-8), Bat rotavirus (mouse rotavirus), Muridbetaherpesvirus 1 (mouse salivary gland virus_mouse cytomegalovirus).

[0017] In another preferred embodiment, the reference genome is obtained from the NCBI genome database.

[0018] In another preferred embodiment, steps (ii) to (iv) are performed using CAMISIM software.

[0019] In another preferred embodiment, the simulated data set in step (ii) is generated by a method selected from the group consisting of de novo simulation, data set simulation, or a combination thereof.

[0020] In another preferred embodiment, the de novo simulation comprises the following steps:

[0021] A1) provide specified taxonomic information, including the name of the taxon and the abundance of the taxon;

[0022] A2) providing the genome classification name of the reference genome in a known database, mapping the taxonomic unit name to the genome classification name of the known database (e.g., NCBI database classification ID), thereby matching the taxonomic unit with the reference genome;

[0023] A3) Based on the novelty level of the reference genome and the corresponding taxonomic unit name, m genomes are extracted from the reference genome to generate the simulated community.

[0024] In another preferred example, the given taxonomic information is a CAMI file or a BIOM file.

[0025] In another preferred example, the species taxonomic unit name refers to the taxonomic operational unit (OTU) taxonomic name.

[0026] In another preferred embodiment, in step A2), the OUT classification name is mapped to the NCBI classification ID.

[0027] In another preferred embodiment, in step A2), the taxon and the genome mapped thereto satisfy the following conditions:

[0028] There is a minimal path between the taxon name and the genome taxon name calculated from the de Bruijn graph, and

[0029] The taxonomic unit name and the genome taxonomic name belong to at least the same family.

[0030] In another preferred embodiment, in step A2), if there is no genome that meets the conditions, the classification unit is assigned to a customized new "random" genome.

[0031] In another preferred embodiment, step A3) comprises the steps of:

[0032] 1) Based on the total number of genomes m, the same number of genomes are extracted from each novelty category (new strain, new species, new genus, new family, new order);

[0033] 2) If there are not enough genomes to be drawn from a certain novelty category, more genomes are drawn from other categories in equal amounts;

[0034] 3) Repeat steps 1) and 2) until m genomes are extracted.

[0035] In another preferred embodiment, the novelty category is determined based on the degree of similarity between the reference genome sequence and currently known viral sequences.

[0036] In another preferred embodiment, the m is 15-40, preferably 20-25.

[0037] In another preferred embodiment, the simulated community further comprises an additional genome simulated by the step, and the additional genome is generated by the following steps:

[0038] a) selecting a specific genome from the reference genome and extracting a strain number from a geometric distribution, wherein the strain number represents the number of related strains created for the specific genome;

[0039] b) repeating this step until the specified total number of strains is reached;

[0040] c) randomly selecting a designated strain number from the generated related strains, thereby obtaining the simulated additional genome.

[0041] In another preferred example, for each reference genome, 20-50 related strains, preferably 30-40 related strains are generated.

[0042] In another preferred embodiment, the genetic distance between each related strain and the starting strain increases in steps of 0.1%, and the total distance is 2%-5%, preferably 3%-4%.

[0043] In another preferred embodiment, the data set simulation comprises the following steps:

[0044] B1) Sampling from the lognormal distribution to create a set of abundance distribution samples;

[0045] B2) generating the simulated community according to the reference genome and the abundance distribution sample set.

[0046] In another preferred embodiment, in step B1), the default data of CAMISIM software is used, and the parameter mu of the lognormal distribution is set to 1, and the parameter sigma is set to 2.

[0047] In another preferred embodiment, in step B1), ≥2 abundance distribution samples, preferably ≥5, and more preferably ≥10, are independently extracted from the lognormal distribution using the same parameter settings.

[0048] In another preferred embodiment, in step B1), for the x+1th abundance distribution sample, a new value is sampled from the lognormal distribution, the new value is added to the value of the xth abundance distribution sample and divided by 2, thereby obtaining the x+1th abundance distribution sample.

[0049] In another preferred embodiment, in step (iii), simulated reads are generated according to specified parameters, wherein the parameters include: read length, insert size, error rate, or a combination thereof.

[0050] In another preferred embodiment, the read length is 50-500 bp, preferably 100-350 bp, and more preferably 150-250 bp.

[0051] In another preferred embodiment, the insert size is 120bp-400bp, preferably 220bp-320bp, and more preferably 270bp-280bp.

[0052] In another preferred embodiment, the error rate is ≤10%, preferably ≤5%, and more preferably ≤1%.

[0053] In another preferred embodiment, in step (iii), sampling is performed in proportion to the designated genome abundance from each genome of the simulated community.

[0054] In another preferred embodiment, the simulated data set contains 10-500 simulated samples, preferably 50-200 simulated samples, and more preferably 100-150 simulated samples.

[0055] In another preferred embodiment, the designated abundance in each sample is different.

[0056] In another preferred embodiment, step (iii) further comprises the step of: generating read simulation for a second time using the same genomic abundance and different average insertion size, and mixing it with the simulated sample obtained previously.

[0057] In another preferred embodiment, in step (iii), the size of each sample in the simulated data set contains 1-10G simulated reads, preferably 3-5G.

[0058] In another preferred embodiment, in step (iii), the generated simulated data set includes for each sample:

[0059] The sequences of simulated reads, such as FASTQ files; and

[0060] Alignment information of simulated reads to a reference genome, such as a BAM file.

[0061] In another preferred embodiment, step (iv) comprises the following steps:

[0062] iva) obtains the position of each read on the reference genome from the read simulation and calculates the coverage of each site;

[0063] ivb) Decomposing the reference genome sequence at zero coverage sites to obtain gold standard contig sequences.

[0064] In another preferred embodiment, step (iv) comprises the following steps:

[0065] ivc) Shuffle and anonymize the sequences of the simulated dataset to obtain the challenge dataset.

[0066] In another preferred embodiment, the method further comprises a step (v) after step (iv): filtering the data according to the bases and sequence quality of the sequencing data of the digital standard; and using sequence alignment software to compare and identify the digital standard with a database built based on a reference genome, and removing sequences that are aligned to non-target viral genomes.

[0067] In another preferred embodiment, the digital standard contains reference genome species and corresponding abundance information contained in each simulated sample.

[0068] In a second aspect of the present invention, a digital standard for metagenomic bioinformatics detection is provided, wherein the digital standard comprises p samples, wherein each of the samples contains a genome sequence derived from a reference genome of n different microorganisms,

[0069] Among them, the p samples have different genome abundances.

[0070] In another preferred embodiment, the microorganism is a virus.

[0071] In another preferred embodiment, the microorganism is n microorganisms selected from the following group: Betapolyomavirusmacacae (simian virus 40), Bluetongue virus (blue tongue virus), Bovine polyomavirus 3 (bovine polyomavirus), Human betaherpesvirus 6A (human herpesvirus-6), Human betaherpesvirus 6B (human herpesvirus-6), Fowl aviadenovirus A (exogenous avian adenovirus type I), Simian retrovirus 8 (monkey retrovirus), Simian foamy virus (monkey foamy virus), Human gammaherpesvirus 4 (human Epstein-Barr virus), Human polyomavirus 12 (human polyomavirus), Human circovirusVS6600022 (human circovirus), Human betaherpesvirus5 (human cytomegalovirus), Humanimmunodeficiency virus 1 (human immunodeficiency virus), Murine respirovirus (Sendai virus), Muridbetaherpesvirus 8 (mouse salivary gland virus_mouse cytomegalovirus), Nipah henipavirus (Nipah virus), Suid alphaherpesvirus 1 (pseudorabies virus), Swinepox virus (swine poxvirus), Porcine reproductive and respiratory syndrome virus (porcine reproductive and respiratory syndrome virus), Suidbetaherpesvirus 2 (porcine cytomegalovirus), Porcine circovirus 2 (porcine circovirus), Humangammaherpesvirus 8 (human herpesvirus-8), Bat rotavirus (mouse rotavirus), Muridbetaherpesvirus 1 (mouse salivary gland virus_mouse cytomegalovirus).

[0072] In another preferred embodiment, the n microorganisms are composed of the following microorganisms: Betapolyomavirusmacacae (simian virus 40), Bluetongue virus (blue tongue virus), Bovine polyomavirus 3 (bovine polyomavirus), Human betaherpesvirus 6A (human herpesvirus-6), Human betaherpesvirus6B (human herpesvirus-6), Fowl aviadenovirus A (exogenous avian adenovirus type I), Simian retrovirus 8 (monkey retrovirus), Simian foamy virus (monkey foamy virus), Human gammaherpesvirus 4 (human Epstein-Barr virus), Human polyomavirus 12 (human polyomavirus), Human circovirusVS6600022 (human circovirus), Human betaherpesvirus5 (human cytomegalovirus), Humanimmunodeficiency virus 1 (human immunodeficiency virus), Murine respirovirus (Sendai virus), Muridbetaherpesvirus 8 (mouse salivary gland virus_mouse cytomegalovirus), Nipah henipavirus (Nipah virus), Suid alphaherpesvirus 1 (pseudorabies virus), Swinepox virus (swine poxvirus), Porcine reproductive and respiratory syndrome virus (porcine reproductive and respiratory syndrome virus), Suidbetaherpesvirus 2 (porcine cytomegalovirus), Porcine circovirus 2 (porcine circovirus), Humangammaherpesvirus 8 (human herpesvirus-8), Bat rotavirus (mouse rotavirus), Muridbetaherpesvirus 1 (mouse salivary gland virus_mouse cytomegalovirus).

[0073] In another preferred embodiment, p is an integer selected from 10-500, preferably 50-200, more preferably 100-150.

[0074] In another preferred embodiment, n is an integer selected from 10-24, preferably 15-24, and more preferably 20-24.

[0075] In another preferred embodiment, the digital standard contains reference genome species and corresponding abundance information contained in each sample.

[0076] In another preferred embodiment, the digital standard is filtered, and the filtering refers to using sequence alignment software to compare and identify the digital standard with a database built based on a reference genome, and removing sequences that are aligned to non-target viral genomes.

[0077] In another preferred embodiment, the digital standard is constructed by the method described in the first aspect of the present invention.

[0078] In a third aspect of the present invention, a device for preparing digital standards for metagenomic bioinformatics detection is provided, the device comprising:

[0079] 1) an input module, wherein the input module is used to input reference genomes of n microorganisms, wherein n is an integer selected from 10-24, preferably 15-24, and more preferably 20-24;

[0080] 2) A simulation module, which is used for:

[0081] generating a simulated community based on the reference genome and taxonomic information;

[0082] Using a read simulator, sampling from each genome of the simulated community to generate simulated samples, and mixing the simulated samples to obtain a simulated data set;

[0083] assembling the simulated data sets to obtain the digital standards; and

[0084] 3) An output module, wherein the output module is used to output the digital standard product.

[0085] In another preferred embodiment, the microorganism is a virus.

[0086] In another preferred embodiment, the microorganism is n microorganisms selected from the following group: Betapolyomavirusmacacae (simian virus 40), Bluetongue virus (blue tongue virus), Bovine polyomavirus 3 (bovine polyomavirus), Human betaherpesvirus 6A (human herpesvirus-6), Human betaherpesvirus 6B (human herpesvirus-6), Fowl aviadenovirus A (exogenous avian adenovirus type I), Simian retrovirus 8 (monkey retrovirus), Simian foamy virus (monkey foamy virus), Human gammaherpesvirus 4 (human Epstein-Barr virus), Human polyomavirus 12 (human polyomavirus), Human circovirusVS6600022 (human circovirus), Human betaherpesvirus5 (human cytomegalovirus), Humanimmunodeficiency virus 1 (human immunodeficiency virus), Murine respirovirus (Sendai virus), Muridbetaherpesvirus 8 (mouse salivary gland virus_mouse cytomegalovirus), Nipah henipavirus (Nipah virus), Suid alphaherpesvirus 1 (pseudorabies virus), Swinepox virus (swine poxvirus), Porcine reproductive and respiratory syndrome virus (porcine reproductive and respiratory syndrome virus), Suidbetaherpesvirus 2 (porcine cytomegalovirus), Porcine circovirus 2 (porcine circovirus), Humangammaherpesvirus 8 (human herpesvirus-8), Bat rotavirus (mouse rotavirus), Muridbetaherpesvirus 1 (mouse salivary gland virus_mouse cytomegalovirus).

[0087] In another preferred embodiment, the n microorganisms are composed of the following microorganisms: Betapolyomavirusmacacae (simian virus 40), Bluetongue virus (blue tongue virus), Bovine polyomavirus 3 (bovine polyomavirus), Human betaherpesvirus 6A (human herpesvirus-6), Human betaherpesvirus6B (human herpesvirus-6), Fowl aviadenovirus A (exogenous avian adenovirus type I), Simian retrovirus 8 (monkey retrovirus), Simian foamy virus (monkey foamy virus), Human gammaherpesvirus 4 (human Epstein-Barr virus), Human polyomavirus 12 (human polyomavirus), Human circovirusVS6600022 (human circovirus), Human betaherpesvirus5 (human cytomegalovirus), Humanimmunodeficiency virus 1 (human immunodeficiency virus), Murine respirovirus (Sendai virus), Muridbetaherpesvirus 8 (mouse salivary gland virus_mouse cytomegalovirus), Nipah henipavirus (Nipah virus), Suid alphaherpesvirus 1 (pseudorabies virus), Swinepox virus (swine poxvirus), Porcine reproductive and respiratory syndrome virus (porcine reproductive and respiratory syndrome virus), Suidbetaherpesvirus 2 (porcine cytomegalovirus), Porcine circovirus 2 (porcine circovirus), Humangammaherpesvirus 8 (human herpesvirus-8), Bat rotavirus (mouse rotavirus), Muridbetaherpesvirus 1 (mouse salivary gland virus_mouse cytomegalovirus).

[0088] In a fourth aspect of the present invention, a method for testing the accuracy of metagenomic biological information detection is provided, comprising the steps of:

[0089] 1) providing a method for detecting the bioinformatics of the metagenome to be tested;

[0090] II) using the detection method to detect each sample in the digital standard product as described in the second aspect of the present invention, thereby obtaining the detection result of the detection method;

[0091] III) comparing the detection results of the detection method with the reference genome species and corresponding abundance information of each sample in the digital standard to determine the accuracy of the detection method.

[0092] In another preferred embodiment, the method includes comparing whether the abundance information of the same virus in a standard is consistent with the detection result of the biological information detection method to be tested.

[0093] In a fifth aspect of the present invention, a device for testing the accuracy of a metagenomic biological information detection method is provided, the device comprising:

[0094] 1) an input module, wherein the input module is used to input a metagenome bioinformatics detection method to be tested;

[0095] II) a calculation module, wherein the calculation module uses the detection method to detect the digital standard product as described in the second aspect of the present invention, thereby obtaining a detection result of the method to be tested;

[0096] III) a comparison module, which compares the detection results of the detection method with the reference genome species and corresponding abundance information of each sample in the digital standard, so as to determine the accuracy of the biological information detection method;

[0097] IV) An output module, which is used to output the accuracy judgment result of the biological information detection method.

[0098] It should be understood that within the scope of the present invention, the above-mentioned technical features of the present invention and the technical features specifically described below (such as embodiments) can be combined with each other to form a new or preferred technical solution. Due to space limitations, they will not be described one by one here. BRIEF DESCRIPTION OF THE DRAWINGS

[0099] The following drawings are used to illustrate specific embodiments of the present invention and are not used to limit the scope of the present invention defined by the claims.

[0100] Figure 1 A flow chart for constructing simulation data for a specific reference genome using CAMISIM is shown.

[0101] Figure 2 An example of the results of using digital standards to validate the accuracy of bioinformatic assay alignments is shown.

[0102] Figure 3 The overlap results of the standards and virus species in the detection results using 59 viruses to construct digital standards are shown.

[0103] Figure 4An example of the results of validating the accuracy of bioinformatic detection alignment using digital standards constructed from 59 viruses is shown. DETAILED DESCRIPTION

[0104] After extensive and in-depth research, the inventors have developed for the first time a digital standard for testing the accuracy of the metagenomic bioinformatics detection process. Specifically, the present invention selects 24 viral genomes of particular concern to the Chinese Pharmacopoeia to construct digital standards that can be used to test the accuracy of the metagenomic bioinformatics detection process, including qualitative and quantitative. The present invention constructed a total of 100 samples, each of which contained genome sequences of 24 different viruses, the abundance was randomly generated, and the abundance distribution of viruses in each sample was different. Through this batch of digital standards, the accuracy of the bioinformatics process based on the metagenomics can be effectively tested. On this basis, the present invention was completed.

[0105] Standards

[0106] The digital standard of the present invention is constructed by the method described in the first aspect of the present invention. Specifically, it is constructed based on a known viral genome selected from the following Table 1:

[0107] Table 1 Virus genome information

[0108]

[0109]

[0110] The size of the viral genome is at least between several K and tens of K. The base difference between the subtype of the virus and the original type may be only 100-200 bases, which accounts for a very small proportion of the entire genome. Since the reads length of the second-generation sequencing is usually 250-300bp, it is difficult to guarantee that sequences containing different bases can be extracted during the random library construction process. The current virus alignment method still relies on the accuracy of sequencing, and generally allows a base mismatch. However, in the distinction between subtypes, a difference of one base may represent different subtypes of viruses, resulting in errors in the final sequencing results. In addition, most sequences between viruses in the current virus library are relatively conservative, and there is no sequence library that can completely distinguish between subtypes of viruses.

[0111] The present invention selects viruses of key concern in the Chinese Pharmacopoeia. In the screening of about 100 viruses of key concern in the Chinese Pharmacopoeia, it is found that since many virus subtypes are included and the genomes contain highly similar virus subtype sequences, the prepared digital standards often contain a large number of similar sequences. Using such digital standards to evaluate the accuracy of bioinformatics detection methods will cause interference, making it difficult to obtain true and reliable evaluation results.

[0112] After a lot of attempts and screening, the present invention selected 24 viruses shown in Table 1. These viruses have a wide range of sources, including bovine, mouse, human and pig sources; and do not contain or basically do not contain similar virus sequences, will not interfere with the test results of digital standards, and can be accurately used to evaluate bioinformatics detection methods; at the same time, these viruses are all key focus viruses of the Chinese Pharmacopoeia, and the digital standards constituted by them have extremely high clinical significance and application value.

[0113] CAMISIM

[0114] The digital standard preparation method of the present invention is performed using the microbial community and metagenomic simulator CAMISIM. CAMISIM software can simulate different microbial abundance profiles, multi-sample time series and differential abundance studies, including real and simulated strain-level diversity, and generate second-generation and third-generation sequencing data.

[0115] CAMISIM allows customization of many properties of generated communities and datasets, such as the overall number of genomes, species diversity, genome abundance distribution, sample size, number of repeats, and sequencing technology used. The sequencing data simulated by the CAMISIM software is highly consistent with the real data. According to the prior art, in two sets of simulated multi-sample data of human and mouse intestinal microbial communities generated by the CAMISIM software, results highly consistent with the real data were observed. At the same time, the effects of different genome divergences, sequencing depths, and read errors on two popular metagenomic sequence assembly software MEGAHIT and metaSPAdes were studied on thousands of small datasets generated by CAMISIM.

[0116] CAMISIM data simulation is divided into three stages:

[0117] 1) Community design, including selection of community members and their genomes, and assignment of their relative abundances; genome selection is based on a truncated geometric distribution, and abundance is based on a lognormal distribution.

[0118] 2) Metagenomic sequencing data simulation;

[0119] 3) Post-processing, including binning and assembly steps.

[0120] Kraken2

[0121] In one embodiment, the digital standard of the present invention can be compared with a known sequence during the preparation process, thereby removing the simulated sequence of the non-target viral genome and improving the accuracy of the final verification. The comparison can be performed using Kraken software.

[0122] Kraken and Kraken2 (iterative version of Kraken) are the most commonly used software in microbiome analysis. Compared with other software with the same function, they are fast and accurate. It is a high-precision metagenomic sequence classification software based on the k-mer algorithm, which can quickly classify sequenced (or simulated) reads into species.

[0123] There are two steps to using Kraken software: preparation (library construction) and identification.

[0124] Step 1: Preparation (library construction)

[0125] ·Build the taxon database corresponding to k-mer (Box1)

[0126] Map the database and index files into memory

[0127] Step 2: Identification

[0128] Cut the sequence to be identified into k-mers

[0129] Align the k-mer to the database to obtain its LCA_taxon (Box 2) and the number of alignments.

[0130] Build the above data into a classification tree, and then calculate the sum of all weights on each root-to-leaf. The largest sum is the classification tree of the sequence.

[0131] Digital Standard Construction Method

[0132] The digital standard construction method of the present invention needs to provide: 1) CAMISIM tool, 2) reference genome of specific microorganisms.

[0133] In one embodiment, the construction method of the present invention includes: 1) installing the CAMISIM tool; 2) downloading the reference genome of a specific microorganism; 3) referring to the CAMISIM tool teaching process, using the reference genome to simulate the standard data (process as shown in Figure 1 4) The simulated data were aligned using Kraken2 (reference), and the sequences aligned to non-target virus genomes were removed.

[0134] In the process of constructing the digital standard of the present invention, a specific amount of simulated data, simulated read length, and insert size are used to obtain better simulated sequence accuracy. The read length used in the construction process of the present invention can be 50-500bp, preferably 100-350bp, and more preferably 150-250bp; the insert size can be 120bp-400bp, preferably 220bp-320bp, and more preferably 270bp-280bp; the size of each sample in the simulated data set can be 1-10G simulated reads, preferably 3-5G. In a preferred embodiment, the construction process of the present invention uses a 150bp read length, the insert size is set to 270, and the data amount of each simulated sample is set to 5G reads.

[0135] Virus detection is performed based on simulated data. Due to sequence similarity issues, there is a probability that the virus will be matched to other non-reference genomes. Therefore, after the digital standard is obtained by simulation, the method of the present invention may also include a step of filtering the digital standard. Specifically, the data is filtered according to the bases and sequence quality of the digital standard sequencing data; and the digital standard is compared and identified with a database built based on the reference genome using sequence alignment software, and the sequence that is matched to the non-target virus genome is removed. The digital standard obtained by the present invention has eliminated sequences that are easily mistakenly matched to sequences that are not in the reference virus genome. When the standard of the present invention is used, basically no mistaken alignment will occur.

[0136] The main advantages of the present invention include:

[0137] 1) The present invention can verify the accuracy of the metagenomic bioinformatics process in the detection of specific viruses. It can help establish the accuracy and versatility of using second-generation sequencing technology for biological product detection, which is conducive to promoting the use of second-generation sequencing in the field of biological products;

[0138] 2) The present invention has developed a set of metagenomic data containing 24 viruses in the Chinese Pharmacopoeia, with known virus types and virus abundance, so it can be used to evaluate the sensitivity and specificity of the metagenomic bioinformatics detection process. At the same time, it can also test the accuracy of the quantitative process.

[0139] 3) The viruses selected in the present invention have a wide range of sources, including bovine, mouse, human and porcine sources; and do not contain or substantially contain similar viral sequences, which significantly reduces the interference of viral subtypes on the test results of digital standards; at the same time, these viruses are all key focus viruses of the Chinese Pharmacopoeia, and the digital standards they constitute have extremely high clinical significance and application value.

[0140] 4) The present invention screened the simulation data volume, simulation read length, and insert size that are more favorable for the construction of digital standards, and obtained better simulation sequence accuracy.

[0141] The present invention will be further described below in conjunction with specific examples. It should be understood that these examples are intended to illustrate the present invention only and are not intended to limit the scope of the present invention. The experimental methods in the following examples where specific conditions are not specified are usually performed under conventional conditions, such as those described in Sambrook et al., Molecular Cloning: A Laboratory Manual (New York: Cold Spring Harbor Laboratory Press, 1989), or under conditions recommended by the manufacturer. Unless otherwise indicated, percentages and parts are weight percentages and weight parts.

[0142] Example 1 Construction of Metagenomic Bioinformatics Detection Standards

[0143] 1.1 Install CAMISIM tool

[0144] Use Git to download version 71a1d86 from https: / / github.com / CAMI-challenge / CAMISIM and install it on the server according to the instructions.

[0145] 1.2 Download reference genomes of specific microorganisms

[0146] The genomes of 24 Chinese Pharmacopoeia focus viruses were downloaded from the NCBI genome database (November 19, 2022 version).

[0147] 1.3 Metagenomic simulation method 1: de novo metagenome simulation

[0148] 1.3.1 Input reference genome

[0149] For de novo metagenomic simulations, all sequences shorter than 1000 base pairs are removed from the provided genome and the sequence is verified to contain only valid characters. The input genome can be a draft genome in fasta format, which may contain ambiguous DNA characters in addition to the bases ACGT, such as "RYWSMKHBVDN". Among them, except for the probability of "N", all other non-base characters are removed.

[0150] 1.3.2 Taxonomic information

[0151] If the input taxonomic information is a taxonomic file in CAMI or BIOM format, the taxonomic units therein are matched with closely related complete genomes (e.g., new genomes) in NCBIRefSeq or other reference files so that the simulated metagenome (genome and its abundance) reflects the input profile as much as possible. This step is completed by matching the metagenome names in the input configuration file. Given the taxonomic name of the taxonomic operational unit (OTU) in the BIOM configuration file, the NCBI taxonomic ID can be inferred. For all reference genomes, a taxonomic ID is also required.

[0152] Next, the lineages between OTUs and genomes are compared and OTUs are mapped to the genome with the minimum path between the OTU taxonomic ID and the genome taxonomic ID, calculated according to the de Bruijn graph. Since the minimum path can be very long, a threshold is set at the family level: OTUs will not be mapped to genomes that do not belong to the same family, i.e., for each mapped genome, they are guaranteed to belong to at least the same family. Typically, this results in some OTUs not being mapped, either because their scientific names do not correspond to NCBI taxonomic IDs or because no complete genome is available for the taxonomic ID. For unmapped OUTs, they can be assigned to "random" genomes in the reference collection to increase complexity with additional options.

[0153] 1.3.3 Community Design

[0154] For de novo community design, a specified number of sequences are selected from all input sequences (e.g., circular elements (i.e., a sequence that can represent a certain microorganism) or genomes) according to certain criteria to generate a specific dataset, and the specified number represents the number of viral sequence species to be simulated. The novelty level is determined based on the similarity of the genome sequence combination to currently known viral sequences. The random selection of genomes for a specific dataset is based on their novelty level and membership in OTUs: in order to increase taxonomic spread, for a specified total number of genomes to be included in a specific dataset, the selection algorithm attempts to extract the same number of genomes from each novelty level (new strain, new species, new genus, new family, new order). If one category does not have enough genomes to come up with the specified number of genomes, more genomes are extracted in equal amounts from other categories if possible. This operation is repeated until the specified number of genomes has been extracted.

[0155] 1.3.4 Additional genome additions

[0156] To increase strain-level diversity in the dataset, many additional genomes can be simulated using sgEvolver12, representing strains related to the reference genome chosen above.

[0157] To do this, a genome was randomly selected and a strain number was drawn from a geometric distribution (p=0.3), which represents the number of related strains created for that particular genome. This process was repeated until the specified total number of strains to be generated was reached. For each genome, sgEvolver simulated 40 strains, each with increasing genetic distance from the previous strain (in steps of 0.1%), up to a total distance of 4%. From the 40 strains, the specified strain number was randomly selected and added to the genome set.

[0158] 1.4 Metagenomic simulation method 2: dataset simulation

[0159] For dataset simulations, both the genome and circular element abundance distributions are required for each dataset.

[0160] For a single sample dataset, the abundance distribution is created by sampling from a lognormal distribution. By default, the log logarithm mu of the logarithmic distribution is set to 1 and sigma is set to 2. A higher sigma will result in a relatively flat and less peaked normal distribution. To reflect differential abundance experiments, two or more abundance samples can be independently drawn from a lognormal distribution using the same parameter settings.

[0161] For consecutive samples in the time series, the abundance of the next sample was calculated by sampling new values ​​from a lognormal distribution, adding these values ​​to the previous value and dividing by two. The absolute abundance of circular elements can be set to 15 times the abundance of microbial genomes to replicate high natural plasmid abundance. For each sample, the taxonomic profile of the higher-ranking taxa was calculated based on the genome abundance values ​​and their taxonomic IDs.

[0162] 1.5 Reading simulation

[0163] For each genome, coverage in the simulated samples is calculated based on the genomic abundance distribution, genome size, and the specified metagenomic sample output size. The selected read simulator then generates reads using the user-specified technology and properties (such as read length, insert size, or error rate (if applicable)), sampling each genome in proportion to the abundance specified in the community. The read length is set to 150bp and the insert size is set to 270bp. The read simulation of the dataset can be run a second time with a different average insert size and the same genomic abundance as before to replicate the use of hybrid technology on the same samples. The size of each sample for each dataset is set to 5G reads.

[0164] The output of this step is, for each sample in the dataset, a FASTQ and BAM file that specifies the position of each read in the input sequence.

[0165] 1.6 Standard product assembly

[0166] Gold standard assemblies were performed using SAMtools for each sample and for all samples per dataset (combined gold standards), respectively, including each sequence position with a coverage of at least 1. Specifically, the SAM file obtained from the read simulation specifies the position of each read on the reference genome and is used with SAMtools to calculate the coverage of each site. The sequences of the reference genome are then decomposed into gold standard contig sequences at zero coverage sites. The contig sequences are then labeled with a unique dataset and sample identification prefix and an increasing index as a suffix. In the last step, the sequences of the simulated dataset are shuffled and anonymized using the unix command "shuf" to create the challenge dataset.

[0167] Example 2 Establishing a Metagenomic Bioinformatics Analysis Process

[0168] Using ab initio simulation, digital standards were constructed according to the method described in Example 1.

[0169] 2.1 Build a self-developed metagenome bioinformatics data analysis process

[0170] a) Use common sequence quality control software such as FastQC or FastP to filter the data based on the bases and sequence quality of the digital standard sequencing data

[0171] b) Download virus sequence alignment software (such as Kraken2, centrifuge, etc.) and use these 24 virus genomes to build a virus database.

[0172] c) Use the above software and preset parameters to perform viral sequence alignment and identification on the sample, and remove sequences that are aligned to non-target viral genomes.

[0173] 2.2 Accuracy test of bioinformatics detection method

[0174] The virus detection results were verified based on the virus types and abundance contained in each digital standard provided by this patent. Figure 2 As shown, the accuracy of the bioinformatics detection method can be judged using the digital standards of the present invention.

[0175] Comparative Example

[0176] Due to the limitations of current sequence alignment methods and virus libraries, it is very difficult to use alignment methods to completely and correctly align subtype viruses. Before trying to use the 24 viruses this time, the inventor used 59 viruses that are of particular concern to the Chinese Pharmacopoeia to prepare and test digital standards. It was found that the standards constructed using 59 viruses produced many false positives. Figure 3As shown, theoretically, the number of detected viruses should be around 59, but after using the bioinformatics process for detection, a total of 136 virus types were detected. Figure 4 The results also showed that some virus types in the standard sample were misdetected or missed.

[0177] After analysis, it was found that this phenomenon may be due to the interference of subtypes in these viruses. The genome sequences in the digital standards are too similar to other sequences that are not in the target virus, resulting in false detection.

[0178] All documents mentioned in the present invention are cited as references in this application, just as each document is cited as reference individually. In addition, it should be understood that after reading the above teachings of the present invention, those skilled in the art can make various changes or modifications to the present invention, and these equivalent forms also fall within the scope defined by the claims attached to this application.

Claims

1. A method for preparing digital standards for metagenomic bioinformatics detection, characterized in that: The following steps are involved: (i) providing n reference genomes of microorganisms, wherein n is an integer selected from 10-24, preferably 15-24, more preferably 20-24; (ii) simulated community design: generating simulated communities based on the reference genome and taxonomic information; (iii) read simulation: using a read simulator to sample from each genome of the simulated community to generate simulated samples, and the simulated samples are mixed to obtain a simulated dataset; (iv) Standard assembly: assembling the simulated data set to obtain the digital standard.

2. The method for preparing a digital standard according to claim 1, characterized in that: The n microorganisms are composed of the following microorganisms: Betapolyomavirus macacae (simian virus 40), Bluetongue virus (blue tongue virus), Bovine polyomavirus 3 (bovine polyomavirus), Human betaherpesvirus6A (human herpesvirus-6), Humanbetaherpesvirus 6B (human herpesvirus-6), Fowl aviadenovirus A (exogenous avian adenovirus type I), Simian retrovirus 8 (monkey retrovirus), Simian foamy virus (monkey foamy virus), Humangammaherpesvirus 4 (human Epstein-Barr virus), Human polyomavirus 12 (human polyomavirus), Human circovirus VS6600022 (human circovirus), Human betaherpesvirus 5 (human cytomegalovirus), Human immunodeficiency virus 1 (human immunodeficiency virus), Murine respirovirus (Sendai virus), Murid betaherpesvirus 8 (mouse salivary gland virus_mouse cytomegalovirus), Nipahhenipavirus (Nipah virus), Suid alphaherpesvirus 1 (pseudorabies virus), Swinepox virus (swine pox virus), Porcine reproductive and respiratory syndrome virus (porcine reproductive and respiratory syndrome virus), Suid betaherpesvirus 2 (porcine cytomegalovirus), Porcine circovirus 2 (porcine circovirus), Human gammaherpesvirus 8 (human herpesvirus-8), Bat rotavirus (mouse rotavirus), Muridbetaherpesvirus 1 (mouse salivary gland virus_mouse cytomegalovirus).

3. The method for preparing a digital standard according to claim 1, characterized in that: The steps (ii) to (iv) are performed using CAMISIM software.

4. The method for preparing a digital standard according to claim 1, characterized in that: The simulated data set comprises 10-500 simulated samples, preferably 50-200 simulated samples, and more preferably 100-150 simulated samples, and the designated abundance in each sample is different.

5. The method for preparing a digital standard according to claim 1, characterized in that: The method further comprises a step (v) after step (iv): filtering the data according to the bases and sequence quality of the sequencing data of the digital standard; and using sequence alignment software to compare and identify the digital standard with a database built based on a reference genome, and removing sequences that are aligned to non-target viral genomes.

6. A digital standard for metagenomic bioinformatics detection, characterized in that: The digital standard comprises p samples, wherein each of the samples contains a genome sequence derived from a reference genome of n different microorganisms, wherein the p samples have different genomic abundances; and p is an integer selected from 10-500, preferably 50-200, more preferably 100-150; n is an integer selected from 10-24, preferably 15-24, more preferably 20-24.

7. The digital standard product according to claim 6, characterized in that: The n microorganisms are composed of the following microorganisms: Betapolyomavirus macacae (simian virus 40), Bluetongue virus (blue tongue virus), Bovinepolyomavirus 3 (bovine polyomavirus), Human betaherpesvirus 6A (human herpesvirus-6), Humanbetaherpesvirus 6B (human herpesvirus-6), Fowl aviadenovirus A (exogenous avian adenovirus type I), Simian retrovirus 8 (monkey retrovirus), Simian foamy virus (monkey foamy virus), Humangammaherpesvirus 4 (human Epstein-Barr virus), Human polyomavirus 12 (human polyomavirus), Human circovirus VS6600022 (human circovirus), Human betaherpesvirus 5 (human cytomegalovirus), Human immunodeficiency virus 1 (human immunodeficiency virus), Murine respirovirus (Sendai virus), Murid betaherpesvirus 8 (mouse salivary gland virus_mouse cytomegalovirus), Nipahhenipavirus (Nipah virus), Suid alphaherpesvirus 1 (pseudorabies virus), Swinepox virus (swine pox virus), Porcine reproductive and respiratory syndrome virus (porcine reproductive and respiratory syndrome virus), Suid betaherpesvirus 2 (porcine cytomegalovirus), Porcine circovirus 2 (porcine circovirus), Human gammaherpesvirus 8 (human herpesvirus-8), Bat rotavirus (mouse rotavirus), Muridbetaherpesvirus 1 (mouse salivary gland virus_mouse cytomegalovirus).

8. A device for preparing digital standards for metagenomic bioinformatics detection, characterized in that: The device comprises: 1) an input module, wherein the input module is used to input reference genomes of n microorganisms, wherein n is an integer selected from 10-24, preferably 15-24, and more preferably 20-24; 2) A simulation module, which is used for: generating a simulated community based on the reference genome and taxonomic information; Using a read simulator, sampling from each genome of the simulated community to generate simulated samples, and mixing the simulated samples to obtain a simulated data set; assembling the simulated data sets to obtain the digital standards; and 3) An output module, wherein the output module is used to output the digital standard product.

9. A method for testing the accuracy of metagenomic biological information detection, characterized in that: Includes steps: 1) providing a method for detecting the bioinformatics of the metagenome to be tested; II) using the detection method to detect each sample in the digital standard product as claimed in claim 6, thereby obtaining the detection result of the detection method; III) comparing the detection results of the detection method with the reference genome species and corresponding abundance information of each sample in the digital standard to determine the accuracy of the detection method.

10. A device for testing the accuracy of a metagenomic bioinformatics detection method, characterized in that: The device includes: 1) an input module, wherein the input module is used to input a metagenome bioinformatics detection method to be tested; II) a calculation module, wherein the calculation module uses the detection method to detect the digital standard product according to claim 6, thereby obtaining a detection result of the method to be tested; III) a comparison module, which compares the detection results of the detection method with the reference genome species and corresponding abundance information of each sample in the digital standard, so as to determine the accuracy of the biological information detection method; IV) An output module, which is used to output the accuracy judgment result of the biological information detection method.

Images