A nanopore multi-targeted microbial identification method, system, device and medium

Abstract

The present invention discloses a nanopore multi-targeted microbial identification method, system, device and medium, which belongs to the field of metagenomics technology based on nanopore sequencing. The method and system of the present invention split chimeric sequences, and can detect more species sequences, especially for certain low-abundance species or species that are prone to chimerism during amplification, the detection sensitivity is significantly improved. Furthermore, the method and system of the present invention corrects the errors of nanopore sequencing to obtain a consistent sequence, further improving the accuracy and credibility of species identification.

Description

Technical Field

[0001] The present invention belongs to the technical field of metagenomics based on nanopore sequencing, and specifically relates to a nanopore multi-targeted microbial identification method, system, equipment and medium. Background Art

[0002] Metagenomics is the study of the structural composition and function of microbial communities (including bacteria, viruses, archaea and fungi, etc.), and is essential for understanding important aspects of environmental and / or human health. Over the past few decades, there has been an important shift in the way microbial communities are explored due to the introduction of sequencing technology. In the late 1980s, scientists realized that the uncultured microbial world was larger than the cultured world, and DNA sequence analysis replaced cultivation to study complex microbial communities. Since then, the emergence of next generation sequencing (NGS) technology has undoubtedly triggered a real revolution in the field of microbiology, with major breakthroughs in the complete characterization of the human gut microbiome or the identification of undiscovered biology. The introduction of third-generation sequencing represents another major turning point in the field, as it opens up the possibility of long-read real-time sequencing.

[0003] The technologies currently dominating this field are single-molecule real-time sequencing (SMRT) and nanopore sequencing (NS), the former commercialized by Pacific Biosciences (PacBio) and the latter by Oxford Nanopore Technologies (ONT). Compared with PacBio, ONT's instruments are smaller, the sequencing throughput is more flexible, and the application scenarios are very wide.

[0004] ONT nanopore sequencing allows single-molecule sequencing based on bioengineered nanopores, which are embedded in a resistive membrane to which a voltage is applied. When a single-stranded DNA / RNA fragment passes through the nanopore, it causes a change in the current through the membrane, and the current is converted into a specific nucleotide sequence using an algorithm based on a recurrent neural network (RNN). Since nanopore sequencing does not require DNA amplification, very long reads (over 2Mb) can be generated, and theoretically there is no limit to the length of the read. Obtaining these ultra-long reads significantly improves the effectiveness of various applications (such as de novo genome assembly, identification of repetitive DNA elements, etc.), which cannot be solved by existing short-read NGS (generally up to 150bp) technology alone. In the identification method of microbial species, long-read sequencing also has inherent advantages in identification specificity and resolution compared to NGS. However, compared with NGS, the error rate of ONT nanopore sequencing is relatively high. Although the sequencing quality is also improving with the update and upgrade of sequencing chips, and it can currently achieve an accuracy of more than 97%, it is still lower than the sequencing quality of NGS Q20 level (99% accuracy), and it is inevitable that some errors in comparison and identification caused by sequencing errors will occur. Furthermore, the sequence of nanopore sequencing may contain some low-complexity sequences (such as GGGGGGG, AGAGAGAG, etc., which are generally repeated sequence units). Current quality control software such as NanoFilt and Filtlong do not process low-complexity sequences, which may also cause subsequent comparison errors. In addition, in terms of species identification, directly using the comparison and identification results of software such as minimap and centrifuge will still have certain false positives due to sequencing errors, incomplete or incorrect annotation of the reference genome, non-specific amplification of primers, etc.

[0005] Targeted metagenomics is an identification method that captures specific genomic regions of multiple species within the detection range. Compared with the metagenomics method, the detection range of targeted metagenomics is smaller, the data volume requirement is lower, it is not easily interfered by the host sequence, it has higher sensitivity and lower cost. Currently, there are two types of methods: probe hybridization capture and multiplex PCR amplification. The probe hybridization capture method is generally more complicated to operate and takes a long time to experiment, while the multiplex PCR amplification method is relatively simpler to operate and takes a short time to experiment, but it has high requirements for the design and verification of multiplex PCR primers. Various parameters of primers (such as Tm value, GC content, primer dimer, etc.) should be considered during the design, and subsequent experiments are required to verify the effect of the primers, and try to ensure that the amplification effect of each primer is relatively balanced and stable. If necessary, the primers with poor effects need to be redesigned and verified. In addition, chimeric sequences may be generated during the PCR process, that is, the sequences of two species are connected into one sequence. The common chimeric processing method is to directly filter out such chimeric sequences, but this may also cause the omission of low-abundance species and cause false negatives. Summary of the invention

[0006] The present invention aims to propose a more rigorous, reasonable and accurate nanopore multiple targeted microbial identification and analysis strategy to address the problems of relatively low sequencing quality, the presence of low-complexity regions, the presence of chimeric sequences, and high false positives in common comparison species identification methods in nanopore multiple targeted sequencing.

[0007] In order to achieve the above object, the technical solution adopted by the present invention is as follows:

[0008] The first aspect of the present invention provides a nanopore multi-targeted microbial identification method, comprising the following steps:

[0009] S1, obtain microbial multiplex targeted nanopore sequencing data;

[0010] S2, aligning the sequencing data with the microbial reference genome to obtain the first reference sequence set,

[0011] S3, each sequencing sequence is used as a query sequence and processed as follows:

[0012] S31, sorting all the first reference sequences aligned with the query sequence from high to low according to the alignment scores, taking the first reference sequence with the highest alignment score as hit0,

[0013] S32, the remaining first reference sequences in the alignment are grouped in descending order according to the alignment scores as follows:

[0014] (1) If the overlap region between the alignment segment between the first reference sequence and the query sequence and the alignment segment between hit0 and the query sequence is greater than or equal to 20% of the length of the alignment segment between the first reference sequence and the query sequence, the first reference sequence and hit0 are grouped together;

[0015] (2) If the overlapped region between the alignment segment between the first reference sequence and the query sequence and the alignment segment between hit0 and the query sequence is less than 20% of the length of the alignment segment between the first reference sequence and the query sequence, the first reference sequence is used as the new hit0, and step S32 is repeated for the remaining aligned first reference sequences.

[0016] S33, extracting the query sequence corresponding to the alignment segment between the first reference sequence and the query sequence with the highest alignment score in each group as a subsequence of the query sequence,

[0017] S4. For each subsequence obtained in step S3, species classification is performed based on the first reference sequence with the highest alignment score in its corresponding group, and the lowest level of species that can be classified is used as the initial species identification result of the subsequence, and the lowest value of the similarity and coverage of the first reference sequence alignment used for species classification is used as the similarity and coverage of the species alignment.

[0018] In some embodiments of the present invention, the microbial multi-target nanopore sequencing data refers to the sequencing data obtained by performing PCR amplification (i.e., multi-target amplification) on a microbial sample using multiple primers, and then performing nanopore sequencing. In some specific embodiments of the present invention, amplification is performed using universal primers or specific primers for bacteria or fungi to obtain the sequence of bacteria or fungi. In some preferred embodiments of the present invention, amplification is performed using universal primers for the 16S rRNA gene of bacteria to obtain sequencing data of the 16S rRNA gene of bacteria, which can be used to identify which bacteria are included in the microbial sample. Of course, amplification can also be performed using universal primers for ITS genes, and further sequencing analysis can be performed to obtain the identification results of fungi. For some special microorganisms, specific primers can also be designed for amplification to improve the sensitivity of detection.

[0019] In some embodiments of the present invention, the microbial reference genome is mainly derived from a public database. For bacteria and fungi, a sequence in the RefSeq targetloci database is selected as the first reference sequence. For sequences amplified by specific primers, the RefSeq database may not have a corresponding first reference sequence, so the public genome of the corresponding species is obtained as the first reference sequence, and the public genome refers to the genome in a peer-reviewed and publicly published article.

[0020] In some embodiments of the present invention, after alignment, each sequencing sequence (read) may correspond to multiple hits, which are called first reference sequences, and the set composed of these first reference sequences is called a first reference sequence set.

[0021] In some embodiments of the present invention, after the alignment is performed in step S2, a step of filtering the alignment results of the sequences is further included, filtering out the results with alignment similarity lower than 85% or alignment coverage lower than 30%, and the filtered sequences are some sequences with low sequencing quality or outside the range of the first reference sequence.

[0022] In some embodiments of the present invention, when all aligned first reference sequences of the query sequence are sorted from high to low according to the alignment scores, there may be multiple first reference sequences with the same alignment score. At this time, the first reference sequences with the same alignment score can be arranged arbitrarily.

[0023] In the present invention, the segment compared with the query sequence refers to the corresponding query sequence fragment compared with the query sequence. After processing in step S3, the chimeric sequence will be split into multiple subsequences corresponding to the grouping. Of course, for a non-chimeric sequence, there may be only one subsequence; further using the first reference sequence with the highest grouping corresponding to each subsequence to perform species classification identification can improve the accuracy of the identification result.

[0024] In some embodiments of the present invention, in step S4, for each subsequence, there are multiple first reference sequences with the highest alignment scores in the corresponding group, and the species corresponding to these multiple first reference sequences are voted, and the classification levels are judged from low to high. If the voting ratio of a certain species classification level is not less than 70%, the species classification of the classification level is used as the preliminary species classification result of the subsequence. For example, for a subsequence, more than 80% of the first reference sequences with the highest alignment scores correspond to Listeria monocytogenes, and the species of the subsequence is identified as Listeria monocytogenes. The alignment similarity and coverage of the first reference sequence corresponding to the species are used as the similarity and coverage of the subsequence aligned to the species.

[0025] In some embodiments of the invention, the method further comprises:

[0026] S5, performing sequence clustering correction on each subsequence obtained in step S3 to generate a representative consensus sequence:

[0027] S51, expressing any alignment position in the form of ref:start-end, wherein ref is the name of the first reference sequence, start indicates the alignment start position, and end indicates the alignment end position,

[0028] S52, select any two comparison positions as the first comparison position loc1 and the second comparison position loc2, respectively expressed as ref1:start1-end1 and ref2:start2-end2,

[0029] If the overlap length of two alignment positions is greater than 20% of the length of the second alignment position, the common alignment interval loc* of loc1 and loc2 is defined as: ref*:start*-end*, where ref* is ref1 or ref2, start* is the minimum value of start1 and start2, and end* is the maximum value of end1 and end2.

[0030] If the overlapping length of the two alignment positions is not greater than 20% of the length of the second alignment position, the two alignment positions are respectively regarded as two different loc*;

[0031] S53, traverse all the comparison positions, and use the same method to determine whether each comparison position overlaps with all loc*. If a comparison position overlaps with a loc*, take the union and update the loc*, otherwise take the comparison position as a new loc*;

[0032] S54, different loc* are obtained in step S52 and S53. For a certain loc*, all sequences belonging to the loc* are clustered into clusters and a consensus sequence of each cluster is generated. The consensus sequence of the cluster containing the most sequences is taken as the representative consensus sequence of the loc*.

[0033] S6, for each subsequence obtained in step S3, perform species identification based on its representative consensus sequence obtained in step S5:

[0034] S61, aligning the representative consensus sequence of the subsequence obtained in step S5 to the NCBI nt database to obtain a second reference sequence set, wherein the highest alignment score is score_max, and selecting a second reference sequence with an alignment score not less than score_max-8;

[0035] S62, using each second reference sequence obtained in step S61 to classify the subsequence into species, and the lowest level species that can be classified is used as the nt verified species identification result of the subsequence, and the lowest value of the similarity and coverage of the second reference sequence used for species classification is used as the similarity and coverage of the subsequence aligned to the species.

[0036] In some embodiments of the present invention, the representative consensus sequence is aligned to the NCBI nt database. Each representative consensus sequence may correspond to multiple hits, which are called second reference sequences. The set composed of these second reference sequences is called a second reference sequence set.

[0037] In some embodiments of the invention, the method further comprises:

[0038] S7, for each subsequence obtained in step S3, an initial species identification result t0 and a verified species identification result t1 are obtained after processing in steps S4 to S6. If the similarity and coverage of t1 are not less than t0, t1 is used as the final species identification result of the subsequence. Otherwise, if the similarity of t0 is not less than 97% and the coverage is not less than 80%, t0 is used as the final species identification result of the subsequence. In other cases, the subsequence is considered to have low quality and no suitable species classification is filtered out.

[0039] In some embodiments of the present invention, in step S1, after obtaining the original offline data of microbial multi-target nanopore sequencing, before step S2, further steps are included, including a sequencing adapter shearing processing step, a low-complexity region splitting step, a low-quality sequence filtering step and / or a short sequence filtering step.

[0040] The low-complexity region splitting step specifically comprises: extracting non-low-complexity regions on both sides of the low-complexity region; the short sequence refers to a sequence with a length of less than 80 bp.

[0041] The different steps in the method of the first aspect of the present invention can be used alone or in combination as independent technical solutions to solve specific technical problems. For example, step S3 can be used alone to perform chimeric sequence splitting.

[0042] A second aspect of the present invention provides a nanopore multi-targeted microbial identification system, comprising:

[0043] A data input module for obtaining microbial multiple targeted nanopore sequencing data;

[0044] A reference genome storage module, used to store microbial reference genomes;

[0045] A comparison module, connected to the data input module and the reference genome storage module respectively, for comparing the sequencing data with the microbial reference genome to obtain a first reference sequence set;

[0046] The chimeric sequence processing module is connected to the alignment module and is used to treat each sequencing sequence as a query sequence and perform the following processing:

[0047] Sort all the first reference sequences on the query sequence according to the alignment score from high to low, and take the first reference sequence with the highest alignment score as hit0.

[0048] The remaining first reference sequences in the alignment are grouped as follows according to the alignment scores from high to low:

[0049] (1) If the overlap region between the alignment segment between the first reference sequence and the query sequence and the alignment segment between hit0 and the query sequence is greater than or equal to 20% of the length of the alignment segment between the first reference sequence and the query sequence, the first reference sequence and hit0 are grouped together;

[0050] (2) If the overlap between the alignment segment between the first reference sequence and the query sequence and the alignment segment between hit0 and the query sequence is less than 20% of the length of the alignment segment between the first reference sequence and the query sequence, then the first reference sequence is used as the new hit0, and this step is repeated for the remaining aligned first reference sequences.

[0051] Extract the query sequence corresponding to the alignment segment between the first reference sequence with the highest alignment score in each group and the query sequence as a subsequence of the query sequence,

[0052] The initial species identification module is connected to the chimeric sequence processing module and is used to classify each subsequence into species based on the first reference sequence with the highest alignment score in its corresponding group, and the lowest level of species that can be classified is used as the initial species identification result of the subsequence, and the lowest value of the similarity and coverage of the first reference sequence used for species classification is used as the similarity and coverage of the subsequence aligned to the species.

[0053] In some embodiments of the invention, the system further comprises:

[0054] A sequence cluster correction module, connected to the chimeric sequence processing module, is used to perform sequence cluster correction on each subsequence to generate a representative consensus sequence:

[0055] Any alignment position is expressed in the form of ref:start-end, where ref is the name of the first reference sequence, start indicates the alignment start position, and end indicates the alignment end position.

[0056] Arbitrary two alignment positions are used as the first alignment position loc1 and the second alignment position loc2, respectively expressed as ref1:start1-end1 and ref2:start2-end2,

[0057] If the overlap length of two alignment positions is greater than 20% of the length of the second alignment position, the common alignment interval loc* of loc1 and loc2 is defined as: ref*:start*-end*, where ref* is ref1 or ref2, start* is the minimum value of start1 and start2, and end* is the maximum value of end1 and end2.

[0058] If the overlapping length of the two alignment positions is not greater than 20% of the length of the second alignment position, the two alignment positions are respectively regarded as two different loc*;

[0059] Traverse all alignment positions and use the same method to determine whether each alignment position overlaps with all loc*. If a certain alignment position overlaps with a certain loc*, take the union and update the loc*. Otherwise, take the alignment position as a new loc* to obtain a different loc*. For a certain loc*, cluster all sequences belonging to the loc* into clusters and generate a consensus sequence for each cluster. Take the consensus sequence of the cluster containing the most sequences as the representative consensus sequence of the loc*.

[0060] nt database storage module, used to store NCBI nt database;

[0061] The nt verification and identification module is connected to the sequence clustering correction module and the nt database storage module respectively, and is used to perform species identification on each subsequence based on its representative consensus sequence obtained by the sequence clustering correction module:

[0062] The representative consensus sequence is aligned to the NCBI nt database to obtain a second reference sequence set, the highest alignment score is score_max, and a second reference sequence with an alignment score not less than score_max-8 is selected;

[0063] The subsequence is classified into species using each second reference sequence, and the lowest level species that can be classified is used as the nt verified species identification result of the subsequence, and the lowest value of the similarity and coverage of the second reference sequence used for species classification is used as the similarity and coverage of the subsequence to the species.

[0064] In some embodiments of the invention, the system further comprises:

[0065] The final identification module is connected to the initial identification module and the nt verification identification module respectively, and is used for final species identification: for each subsequence, there is a corresponding initial species identification result t0 and nt verification species identification result t1. If the similarity and coverage of t1 are not less than t0, t1 is used as the final species identification result of the subsequence. Otherwise, if the similarity of t0 is not less than 97% and the coverage is not less than 80%, t0 is used as the final species identification result of the subsequence. In other cases, it is considered that the subsequence is of low quality and no suitable species classification is filtered out.

[0066] In some embodiments of the present invention, before the data input module and the comparison module, pre-processing is further included, including sequencing adapter shearing, low-complexity region splitting, low-quality sequence filtering and / or short sequence,

[0067] The low-complexity region splitting is specifically: extracting non-low-complexity regions on both sides of the low-complexity region; the short sequence refers to a sequence with a length of less than 80 bp.

[0068] Similarly, different modules in the system of the second aspect of the present invention can be used alone or in combination as independent technical solutions to solve specific technical problems. For example, the chimeric sequence processing module can be used alone to split the chimeric sequence.

[0069] The third aspect of the present invention provides a computer device, comprising a memory for storing a computer program; and a processor for implementing the steps of a nanopore multi-targeted microbial identification method as described in any one of the first aspects of the present invention when executing the computer program.

[0070] The fourth aspect of the present invention provides a computer-readable storage medium having a computer program stored thereon, and when the computer program is executed by a processor, the steps of a nanopore multi-targeted microbial identification method as described in any one of the first aspect of the present invention are implemented.

[0071] Beneficial Effects of the Invention

[0072] Compared with the prior art, the present invention has the following beneficial effects:

[0073] The method and system of the present invention can process low-complexity sequences generated by nanopore sequencing and reduce subsequent comparison and identification errors.

[0074] In the prior art, chimeric sequences are directly removed, while the method and system of the present invention split the chimeric sequences, which can detect more species sequences, especially for certain low-abundance species or species that are prone to chimerism during amplification, and the detection sensitivity is significantly improved.

[0075] The method and system of the present invention correct the errors in nanopore sequencing to obtain a consensus sequence, further improving the accuracy and credibility of identification.

[0076] The method and system of the present invention further verify the nt library of the consensus sequence for the preliminary species identification results, which can reduce the false positive effects caused by non-specific amplification of primers, incomplete reference sequences in the target database, annotation errors, etc., and make the results more accurate. BRIEF DESCRIPTION OF THE DRAWINGS

[0077] Figure 1 A flow chart of the nanopore multi-targeted microbial identification method according to Example 1 of the present invention is shown.

[0078] Figure 2 A schematic diagram of a nanopore multi-targeted microbial identification system according to Example 2 of the present invention is shown. DETAILED DESCRIPTION

[0079] In order to make the technical problems, technical solutions and beneficial effects solved by the present invention more clearly understood, the present invention is further described in detail below in conjunction with embodiments.

[0080] The following examples are used to demonstrate preferred embodiments of the present invention. It will be appreciated by those skilled in the art that the techniques disclosed in the following examples represent techniques discovered by the inventors that can be used to implement the present invention and therefore can be considered as preferred embodiments of the present invention. However, it will be appreciated by those skilled in the art based on this specification that many modifications may be made to the specific embodiments disclosed herein and still achieve the same or similar results without departing from the spirit or scope of the present invention.

[0081] Unless defined otherwise, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention belongs, and the disclosure and materials cited therein are hereby incorporated by reference.

[0082] Those skilled in the art will recognize, or be able to ascertain using no more than routine experimentation, many technical equivalents to the specific embodiments of the invention described herein. Such equivalents are intended to be encompassed by the claims.

[0083] The experimental methods in the following examples are conventional methods unless otherwise specified. The instruments and equipment used in the following examples are conventional laboratory instruments and equipment unless otherwise specified; the experimental materials used in the following examples are purchased from conventional biochemical reagent stores unless otherwise specified.

[0084] Example 1 Nanopore multi-targeted microbial identification method

[0085] This embodiment provides a process for implementing a nanopore multi-targeted microbial identification method, such as Figure 1 As shown:

[0086] (1) Data preprocessing

[0087] The nanopore sequencing data was preprocessed, including the processing of sequencing adapters, low-complexity regions, and low-quality sequences, and finally sequences of sufficient length were retained.

[0088] 1) Shearing of sequencing adapters

[0089] Use Porechop software to search for sequencing adapters in the sequence through the adapter sequence library inside Porechop. The software will first select a part of the sequencing sequences (10,000 by default), and the adapter matching similarity is set to a minimum of 80%. That is, if one of the selected sequences matches a certain adapter in the adapter library, it is considered that the adapter exists.

[0090] If the adapter matches the end of the sequence (the default is within 150bp at both ends of the sequence), if the length of the matching part is long enough (not less than 20bp), the adapter is directly cut to the end of the sequence; when the adapter does not match the end of the sequence, that is, it matches the middle of the sequence, and the minimum matching similarity is not less than 80%, the sequence is split into two segments. For the split sequence fragments, the sequence fragments with a length of not less than 80bp are retained, otherwise they are filtered out, and the other parameters of Porechop remain default.

[0091] 2) Processing of low-complexity areas

[0092] First, use the fastx_mask function of vsearch to identify low-complexity regions. The recognition algorithm selects dust by default. The bases in the identified low-complexity regions will be marked as N. If the proportion of non-low-complexity regions is not less than 70%, the sequence will be retained for subsequent processing. Otherwise, the low-complexity region is considered to be too high and is filtered out.

[0093] For the retained sequence, if there are several low-complexity regions marked as continuous N in the sequence, further processing is required to remove the low-complexity regions as much as possible. The specific method is: for all low-complexity regions with a length of not less than 10bp, extract the non-low-complexity regions on both sides of the region, which is equivalent to using the low-complexity regions to split the sequence into multiple non-low-complexity subsequences, and retain the subsequences with longer length (not less than 50bp).

[0094] 3) Processing low-quality sequences and filtering sequence length

[0095] Filtlong software was used to retain sequences with an average sequencing quality score of no less than 70 points and a length of no less than 80 bp.

[0096] (2) Reference genome alignment

[0097] Use minimap2 to align the preprocessed sequence to the reference genome sequence containing the target region within the detection range, and also to the host reference genome to obtain the position of the sequence on the genome. The minimap2 alignment mode parameter is "-x map-ont" to support Nanopore 3rd generation nanopore data. For each sequencing sequence (read), there will be a primary alignment result by default, and a maximum of 5 secondary alignment results will be retained by default, because there may be a possibility that a sequencing sequence is aligned to multiple similar reference sequences.

[0098] The construction method of the reference genome within the targeted range here: For the reference sequence of the 16S and ITS universal primer amplification range, select the sequence of RefSeq targetloci (https: / / www.ncbi.nlm.nih.gov / refseq / targetedloci / ) as the reference sequence. For the reference sequence of the specific primer amplification range, first select the representative genome in the RefSeq database as the reference sequence. If there is no representative genome sequence in the RefSeq database, obtain the public genome of the corresponding species as the first reference sequence. The public genome refers to the genome in the peer-reviewed and publicly published article. In this embodiment, the inventor searches for the Latin name of the corresponding species in GenBank (https: / / www.ncbi.nlm.nih.gov / genbank / ), clicks on a record to display the specific information of the sequence, and checks whether there is a PubMed document number. If the corresponding document is peer-reviewed and publicly published, it is considered that the reference sequence is of qualified quality and can be selected for inclusion.

[0099] (3) Chimeric sequence processing and preliminary species classification

[0100] First, the sequence alignment results are preliminarily filtered to filter out results with alignment similarities (identities) lower than 85% and alignment coverage (query cover, the percentage of sequence alignment area in the total sequence length) lower than 30%. The filtered sequences are generally those with low sequencing quality or outside the reference sequence range.

[0101] All reference sequence alignment results (also called hits) of each sequencing sequence (also called query) are sorted from high to low according to the alignment score, and then the hits are grouped according to the query sequence alignment segment corresponding to each hit. The hit with the highest alignment score is first taken out and called hit0. If the overlapping area of ​​the query sequence alignment segment of the current hit and the query sequence alignment segment of hit0 is greater than or equal to 20% of the length of the query sequence alignment segment of the current hit, the current hit and hit0 are grouped together. Otherwise, the current hit is used as the new hit0 and the remaining hits are grouped in the same way.

[0102] After the above operation, a query sequence is split into multiple (or only one) groups, and the sequence of the alignment segment of the query sequence corresponding to the hit with the highest alignment score in each group is extracted as a subsequence. Thus, a query sequence is split into multiple (or only one) subsequences. This method can split chimeric sequences.

[0103] For the species classification of a subsequence, first take out the hits with the highest alignment score in the corresponding group, and vote on the species classification of these hits with the highest scores, referring to the classification level of NCBI (domain, kingdom, phylum, class, order, family, genus, species), and judge the classification level from low to high (species is the lowest and domain is the highest). If the voting ratio of a species classification level is not less than 70%, the species classification of this classification level is used as the preliminary species classification result of the subsequence, and the alignment similarity and coverage are recorded according to the hits used for species inference. The alignment similarity and coverage here are the lowest values ​​of the similarity and coverage of these hits.

[0104] (4) Sequence clustering correction to generate representative consensus sequences

[0105] For each subsequence obtained by splitting in step (3), according to the preliminary alignment position coordinates of the subsequence, the sequences with alignment positions in the same region are grouped into one group. The specific steps are as follows:

[0106] For any two alignment positions loc1 and loc2, they can be written in the form of ref:start-end, where ref is the name of a reference sequence, start indicates the alignment start position, and end indicates the alignment end position. Here, loc1 and loc2 can also be expressed as ref1:start1-end1 and ref2:start2-end2 respectively. If ref1 and ref2 are the same, and the intervals of loc1 and loc2 overlap and the overlapping length is greater than 20% of the interval length of loc2 (end2-start2+1), then the common alignment interval loc* of loc1 and loc2 is defined as: ref*:start*-end*, where ref* is equal to ref1 or ref2, start* is the minimum value of start1 and start2, and end* is the maximum value of end1 and end2, which is equivalent to taking lo The union of c1 and loc2, record ref*:start*-end*, and then traverse the alignment position loc_i of each sequence. In the same way, loc_i is used to determine whether it overlaps with all the recorded common positions loc*. If it overlaps, take the union and update loc*, otherwise record loc_i as a new loc*. After traversing all sequences, a group of common positions loc* will be obtained, and a certain loc* is called a group. For all sequences belonging to a group, use the clustering function of vs earch (--cluster_size parameter) to further cluster the sequences in the group according to the minimum similarity (--id parameter) 95% and generate the consensus sequence of each cluster (--consout output), and finally take the consensus sequence of the cluster with the highest abundance, that is, the one containing the most sequences, as the representative consensus sequence of the entire group for subsequent nt library comparison verification.

[0107] (5) Consensus sequence comparison with nt library for verification

[0108] The representative consensus sequences generated above were aligned to the NCBI nt database using NCBI blast ( https: / / ftp.ncbi.nlm.nih.gov / blast / db / ), the minimum similarity of the alignment is 95%, the minimum coverage is 70%, and the top 100 alignment hits with the highest alignment scores are displayed. Assuming that the highest alignment score of these hits is score_max, hits with an alignment score not less than score_max-8 are screened for subsequent species inference, that is, a certain alignment score deviation is allowed and a certain sequence diversity is tolerated.

[0109] The species classification of these screened hits is voted on, referring to the NCBI classification hierarchy (domain, kingdom, phylum, class, order, family, genus, species), and judged from low to high classification levels (species is the lowest and domain is the highest). If the voting ratio of a certain species classification level is not less than 70%, the species classification of this classification level is used to verify the classification result of the nt library of the subsequence, and the final alignment similarity and coverage of the consensus sequence are recorded according to the hits used for species inference. The alignment similarity and coverage here are defined as the minimum values ​​of the similarity and coverage of these hits alignments.

[0110] (6) Final species classification and abundance calculation

[0111] For each subsequence obtained by splitting in step (3), there is a preliminary species classification t0 and a corresponding nt library verification classification t1 of the consensus sequence. If the similarity and coverage of t1 are both greater than or equal to the corresponding values ​​of t0, t1 is used as the final species classification of the subsequence. Otherwise, if the similarity of t0 is greater than or equal to 97% and the coverage is greater than or equal to 80%, t0 is used as the final species classification of the subsequence. In other cases, the subsequence is considered to be of low quality and has no suitable species classification and is filtered out. The total number of subsequences with non-host species classification is recorded as n. According to the NCBI classification hierarchy, the total number of subsequences of a species is recorded as m, and the abundance of the species is defined as m / n.

[0112] Example 2 Nanopore multi-targeted microbial identification system

[0113] This embodiment provides a corresponding system based on the method of embodiment 1, as shown in the schematic diagram Figure 2 As shown, including:

[0114] (1) Data input module

[0115] Used to obtain microbial multiplex targeted nanopore sequencing data.

[0116] (2) Reference genome storage module

[0117] Used to store microbial reference genomes.

[0118] (3) Comparison module

[0119] They are respectively connected to the data input module and the reference genome storage module to compare the sequencing data with the microbial reference genome to obtain a first reference sequence set.

[0120] (4) Chimeric sequence processing module

[0121] Connected to the alignment module, it is used to treat each sequencing sequence as a query sequence and perform the following processing:

[0122] Sort all the first reference sequences on the query sequence according to the alignment score from high to low, and take the first reference sequence with the highest alignment score as hit0.

[0123] The remaining first reference sequences in the alignment are grouped as follows according to the alignment scores from high to low:

[0124] (1) If the overlap region between the alignment segment between the first reference sequence and the query sequence and the alignment segment between hit0 and the query sequence is greater than or equal to 20% of the length of the alignment segment between the first reference sequence and the query sequence, the first reference sequence and hit0 are grouped together;

[0125] (2) If the overlap between the alignment segment between the first reference sequence and the query sequence and the alignment segment between hit0 and the query sequence is less than 20% of the length of the alignment segment between the first reference sequence and the query sequence, then the first reference sequence is used as the new hit0, and this step is repeated for the remaining aligned first reference sequences.

[0126] The query sequence corresponding to the alignment segment between the first reference sequence with the highest alignment score in each group and the query sequence is extracted as a subsequence of the query sequence.

[0127] (5) Initial species identification module

[0128] The method is connected to the chimeric sequence processing module and is used to classify each subsequence into species based on the first reference sequence with the highest alignment score in the corresponding group obtained by the chimeric sequence processing module, and the lowest level of species that can be classified is used as the initial species identification result of the subsequence, and the lowest value of the similarity and coverage of the first reference sequence used for species classification is used as the similarity and coverage of the subsequence aligned to the species.

[0129] (6) Sequence clustering correction module

[0130] Connected with the chimeric sequence processing module, it is used to perform sequence clustering correction on each subsequence to generate a representative consensus sequence:

[0131] Any alignment position is expressed in the form of ref:start-end, where ref is the name of the first reference sequence, start indicates the alignment start position, and end indicates the alignment end position.

[0132] Arbitrary two alignment positions are used as the first alignment position loc1 and the second alignment position loc2, respectively expressed as ref1:start1-end1 and ref2:start2-end2,

[0133] If the overlap length of two alignment positions is greater than 20% of the length of the second alignment position, the common alignment interval loc* of loc1 and loc2 is defined as: ref*:start*-end*, where ref* is ref1 or ref2, start* is the minimum value of start1 and start2, and end* is the maximum value of end1 and end2.

[0134] If the overlapping length of the two alignment positions is not greater than 20% of the length of the second alignment position, the two alignment positions are respectively regarded as two different loc*;

[0135] Traverse all alignment positions and use the same method to determine whether each alignment position overlaps with all loc*. If a certain alignment position overlaps with a certain loc*, take the union and update the loc*. Otherwise, take the alignment position as a new loc* to obtain a different loc*. For a certain loc*, cluster all sequences belonging to the loc* into clusters and generate a consensus sequence for each cluster. Take the consensus sequence of the cluster containing the most sequences as the representative consensus sequence of the loc*.

[0136] (7) nt database storage module

[0137] Used to store NCBI nt database.

[0138] (8)nt verification and identification module

[0139] They are respectively connected to the sequence clustering correction module and the nt database storage module, and are used to perform species identification on each subsequence based on its representative consensus sequence obtained by the sequence clustering correction module:

[0140] The representative consensus sequence is aligned to the NCBI nt database to obtain a second reference sequence set, the highest alignment score is score_max, and a second reference sequence with an alignment score not less than score_max-8 is selected;

[0141] The subsequence is classified into species using each second reference sequence, and the lowest level species that can be classified is used as the nt verified species identification result of the subsequence, and the lowest value of the similarity and coverage of the second reference sequence used for species classification is used as the similarity and coverage of the subsequence to the species.

[0142] (9) Final identification module

[0143] They are respectively connected to the initial identification module and the nt verification identification module for final species identification: for each subsequence, there is a corresponding initial species identification result t0 and nt verification species identification result t1. If the similarity and coverage of t1 are not less than t0, t1 is used as the final species identification result of the subsequence. Otherwise, if the similarity of t0 is not less than 97% and the coverage is not less than 80%, t0 is used as the final species identification result of the subsequence. In other cases, the subsequence is considered to have low quality and no suitable species classification is filtered out.

[0144] Example 3 Application of Nanopore Multi-targeted Microbial Identification Method or System

[0145] In this example, a ZymoBIOMICSD6300 standard bacterial solution sample was selected for detection, DNA of the bacterial solution was extracted, two pairs of universal primers for bacterial 16S rRNA plus specific primers for some species were used for amplification, and then nanopore third-generation sequencing was performed. The sequencing data was analyzed according to the method flow of Example 1 or the system of Example 2 to determine whether the species detection was consistent.

[0146] Among them, the D6300 species covered by the amplification primers are shown in Table 1.

[0147] Table 1 Primers corresponding to each species

[0148] Species Latin name Species Chinese name Corresponding amplification primers Pseudomonas aeruginosa Pseudomonas aeruginosa 16S universal primer Escherichia coli Escherichia coli 16S universal primer Salmonella enterica Salmonella enterica 16S universal primer Limosilactobacillus fermentum Lactobacillus fermentum 16S universal primer Enterococcus faecalis Enterococcus faecalis 16S universal primer Staphylococcus aureus Staphylococcus aureus Specific primers Listeria monocytogenes Listeria monocytogenes 16S universal primer Bacillus subtilis Bacillus subtilis 16S universal primer

[0149] Among them, the specific primer sequences are shown in Table 2:

[0150] Table 2 Primer sequences

[0151] Primer name Forward Primer Reverse primer 16S universal primer (V3-V4 region) GTGCCAGCMGCCGCGGTAA GACTACCAGGGTATCTAATC 16S universal primer (V5-V6 region) TAAAACTYAAAKGAATTGACGGG CGGTTACCTTGTTACGACTT Staphylococcus aureus specific primers TCGCTTGCTATGATTGTGG ACATACGCCAATGTTCTACC

[0152] The main results are described as follows:

[0153] (1) Low-complexity sequence processing in data preprocessing

[0154] The original sequence contains the following sequences, taking the sequence with id 0cb8cba8-8f63-4806-b6ab-1d8e551fe7b1 as an example:

[0155] @0cb8cba8-8f63-4806-b6ab-1d8e551fe7b1

[0156] AGGGGTTGGACAGGAAGGACACCATGCCGCCGACGGCCGAGATCACGGCCCCGGCGCCCAAGGACCACCTTGCCGACGGCGGCGCCCACGTCGCCGAGGCCCTGGAAGAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAATGTACTTCAGGATCGGTTACATGTGCTAAGGTTAACTTTGTTGTTGACTCGACGGTAGCAGCACCTGTCTTTGGAGCGAAGCGTTTTCGACCCGGCTTTTTAACTTCTGCGGGTCTGAAGGTCACCACGCGCCGTGCTGTAATGGGAATATCCTCGCCCGTTTTCGGGTTACGTCCCGGGCGTTGATTCTTATCACGCAGATCGAAGTTACCAAAACCAGAGAGTTTCACCTGTTCGCCGTTTTCCAGAGCGCGACGGATCTCTTCGAAAAACAGTTCAACCGGTTCTTTGTGT

[0157] After low-complexity processing, it can be transformed into the following two sequences:

[0158] @0cb8cba8-8f63-4806-b6ab-1d8e551fe7b1_1

[0159] AGGGGTTGGACAGGAAGGACACCATGCCGCCGACGGCCGAGATCACGGCCCCGGCGCCCAAGGACCACCTTGCCGACGGCGGCGCCCACGTCGCCGAGGCCCTGGAAG

[0160] @0cb8cba8-8f63-4806-b6ab-1d8e551fe7b1_2

[0161] TGTACTTCAGGATCGGTTACATGTGCTAAGGTTAACTTTGTTGTTGACTCGACGGTAGCAGCACCTGTCTTTGGAGCGAAGCGTTTTCGACCCGGCTTTTTAACTTCTGCGGGTCTGAAGGTCACCACGCGCCGTGCTGTAATGGGA ATATCCTCGCCCGTTTTCGGGTTACGTCCCGGGCGTTGATTCTTATCACCGCAGATCGAAGTTACCAAAACCAGAGAGTTTCACCTGTTCGCCGTTTTCCAGAGCGCGACGGATCTCTTCGAAAAACAGTTCAACCGGTTCTTTGTGT

[0162] It can be seen that the low-complexity sequence processing step can effectively identify low-complexity regions and perform splitting processing.

[0163] (2) Alignment and chimeric sequence processing

[0164] The method of Example 1 and the system of Example 2 can accurately split the chimeric sequence. Taking the sequence 01073e8b-ad99-46b5-9cec-c5ab5854ccf0 as an example, its minimap alignment result is shown in Table 3.

[0165] Table 3 Alignment results before chimeric sequence processing

[0166]

[0167] Here, the sequence itself is 636 bp in length, and alignment revealed that it was divided into two segments (subsequences), which were aligned to Staphylococcus aureus (NZ_OX344719.1) and Escherichia coli (NZ_OU701449.1), respectively.

[0168] After chimera processing, the sequence was split into two segments, and the results are shown in Table 4:

[0169] Table 4 Alignment results after chimeric sequence processing

[0170]

[0171] It can be seen that the method of Example 1 or the system of Example 2 can effectively split chimeric sequences and improve the sensitivity of species detection.

[0172] (3) Species identification results

[0173] The final species identification results are shown in Table 5:

[0174] Table 5 Species identification results

[0175]

[0176]

[0177] As can be seen from Table 5, for low-abundance species, such as Enterococcus faecalis, only 357 sequences can be detected by directly removing chimeric sequences in the prior art, while 777 sequences can be detected by the chimeric sequence splitting method of the present invention, an increase of 117%, and the detection sensitivity is significantly improved. For species that are prone to chimerism during amplification, such as Staphylococcus aureus, the number of sequence detections is also significantly increased by the chimeric sequence splitting method of the present invention.

[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 nanopore multi-targeted microbial identification method, characterized in that: The following steps are involved: S1, obtain microbial multiplex targeted nanopore sequencing data; S2, aligning the sequencing data with the microbial reference genome to obtain the first reference sequence set, S3, each sequencing sequence is used as a query sequence and processed as follows: S31, sorting all the first reference sequences aligned with the query sequence from high to low according to the alignment scores, taking the first reference sequence with the highest alignment score as hit0, S32, the remaining first reference sequences in the alignment are grouped in descending order according to the alignment scores as follows: (1) If the overlap region between the alignment segment between the first reference sequence and the query sequence and the alignment segment between hit0 and the query sequence is greater than or equal to 20% of the length of the alignment segment between the first reference sequence and the query sequence, the first reference sequence and hit0 are grouped together; (2) If the overlapped region between the alignment segment between the first reference sequence and the query sequence and the alignment segment between hit0 and the query sequence is less than 20% of the length of the alignment segment between the first reference sequence and the query sequence, the first reference sequence is used as the new hit0, and step S32 is repeated for the remaining aligned first reference sequences. S33, extracting the query sequence corresponding to the alignment segment between the first reference sequence and the query sequence with the highest alignment score in each group as a subsequence of the query sequence, S4. For each subsequence obtained in step S3, species classification is performed based on the first reference sequence with the highest alignment score in its corresponding group, and the lowest level of species that can be classified is used as the initial species identification result of the subsequence. The lowest value of the similarity and coverage of the first reference sequence used for species classification is used as the similarity and coverage of the subsequence aligned to the species.

2. A nanopore multi-targeted microbial identification method according to claim 1, characterized in that: Further including: S5, performing sequence clustering correction on each subsequence obtained in step S3 to generate a representative consensus sequence: S51, expressing any alignment position in the form of ref:start-end, wherein ref is the name of the first reference sequence, start indicates the alignment start position, and end indicates the alignment end position, S52, select any two comparison positions as the first comparison position loc1 and the second comparison position loc2, respectively expressed as ref1:start1-end1 and ref2:start2-end2, If the overlap length of two alignment positions is greater than 20% of the length of the second alignment position, the common alignment interval loc* of loc1 and loc2 is defined as: ref*:start*-end*, where ref* is ref1 or ref2, start* is the minimum value of start1 and start2, and end* is the maximum value of end1 and end2. If the overlapping length of the two alignment positions is not greater than 20% of the length of the second alignment position, the two alignment positions are respectively regarded as two different loc*; S53, traverse all the comparison positions, and use the same method to determine whether each comparison position overlaps with all loc*. If a comparison position overlaps with a loc*, take the union and update the loc*, otherwise take the comparison position as a new loc*; S54, different loc* are obtained in step S52 and S53. For a certain loc*, all sequences belonging to the loc* are clustered into clusters and a consensus sequence of each cluster is generated. The consensus sequence of the cluster containing the most sequences is taken as the representative consensus sequence of the loc*. S6, for each subsequence obtained in step S3, perform species identification based on its representative consensus sequence obtained in step S5: S61, aligning the representative consensus sequence of the subsequence obtained in step S5 to the NCBI nt database to obtain a second reference sequence set, wherein the highest alignment score is score_max, and selecting a second reference sequence with an alignment score not less than score_max-8; S62, using each second reference sequence obtained in step S61 to classify the subsequence into species, and the lowest level species that can be classified is used as the nt verified species identification result of the subsequence, and the lowest value of the similarity and coverage of the second reference sequence used for species classification is used as the similarity and coverage of the subsequence aligned to the species.

3. A nanopore multi-targeted microbial identification method according to claim 2, characterized in that: Further including: S7, for each subsequence obtained in step S3, an initial species identification result t0 and a verified species identification result t1 are obtained after processing in steps S4 to S6. If the similarity and coverage of t1 are not less than t0, t1 is used as the final species identification result of the subsequence. Otherwise, if the similarity of t0 is not less than 97% and the coverage is not less than 80%, t0 is used as the final species identification result of the subsequence. In other cases, the subsequence is considered to have low quality and no suitable species classification is filtered out.

4. A nanopore multi-targeted microbial identification method according to any one of claims 1 to 3, characterized in that: In step S1, after obtaining the original data of microbial multi-target nanopore sequencing, before step S2, the steps further include a sequencing adapter shearing processing step, a low-complexity region splitting step, a low-quality sequence filtering step and / or a short sequence filtering step. The low-complexity region splitting step specifically comprises: extracting non-low-complexity regions on both sides of the low-complexity region; the short sequence refers to a sequence with a length of less than 80 bp.

5. A nanopore multi-targeted microbial identification system, characterized in that: include: A data input module for obtaining microbial multiple targeted nanopore sequencing data; A reference genome storage module, used to store microbial reference genomes; A comparison module, connected to the data input module and the reference genome storage module respectively, for comparing the sequencing data with the microbial reference genome to obtain a first reference sequence set; The chimeric sequence processing module is connected to the alignment module and is used to treat each sequencing sequence as a query sequence and perform the following processing: Sort all the first reference sequences on the query sequence according to the alignment score from high to low, and take the first reference sequence with the highest alignment score as hit0. The remaining first reference sequences in the alignment are grouped as follows according to the alignment scores from high to low: (1) If the overlap region between the alignment segment between the first reference sequence and the query sequence and the alignment segment between hit0 and the query sequence is greater than or equal to 20% of the length of the alignment segment between the first reference sequence and the query sequence, the first reference sequence and hit0 are grouped together; (2) If the overlap between the alignment segment of the first reference sequence and the query sequence and the hit0 and query sequence segment is less than 20% of the length of the alignment segment of the first reference sequence and the query sequence, then the reference sequence is used as the new hit0, and this step is repeated for the remaining aligned first reference sequences. Extract the query sequence corresponding to the alignment segment between the first reference sequence with the highest alignment score in each group and the query sequence as a subsequence of the query sequence, An initial species identification module is connected to the chimeric sequence processing module and is used to classify each subsequence into species based on the first reference sequence with the highest alignment score in its corresponding group, and the lowest level of species that can be classified is used as the initial species identification result of the subsequence, and the lowest value of the similarity and coverage of the first reference sequence used for species classification is used as the similarity and coverage of the subsequence aligned to the species.

6. A nanopore multi-targeted microbial identification system according to claim 5, characterized in that: Also includes: A sequence cluster correction module, connected to the chimeric sequence processing module, is used to perform sequence cluster correction on each subsequence to generate a representative consensus sequence: Any alignment position is expressed in the form of ref:start-end, where ref is the name of the first reference sequence, start indicates the alignment start position, and end indicates the alignment end position. Arbitrary two alignment positions are used as the first alignment position loc1 and the second alignment position loc2, respectively expressed as ref1:start1-end1 and ref2:start2-end2, If the overlap length of two alignment positions is greater than 20% of the length of the second alignment position, the common alignment interval loc* of loc1 and loc2 is defined as: ref*:start*-end*, where ref* is ref1 or ref2, start* is the minimum value of start1 and start2, and end* is the maximum value of end1 and end2. If the overlapping length of the two alignment positions is not greater than 20% of the length of the second alignment position, the two alignment positions are respectively regarded as two different loc*; Traverse all alignment positions and use the same method to determine whether each alignment position overlaps with all loc*. If a certain alignment position overlaps with a certain loc*, take the union and update the loc*. Otherwise, take the alignment position as a new loc* to obtain a different loc*. For a certain loc*, cluster all sequences belonging to the loc* into clusters and generate a consensus sequence for each cluster. Take the consensus sequence of the cluster containing the most sequences as the representative consensus sequence of the loc*. nt database storage module, used to store NCBI nt database; The nt verification and identification module is connected to the sequence clustering correction module and the nt database storage module respectively, and is used to perform species identification on each subsequence based on its representative consensus sequence obtained by the sequence clustering correction module: The representative consensus sequence is aligned to the NCBI nt database to obtain a second reference sequence set, the highest alignment score is score_max, and a second reference sequence with an alignment score not less than score_max-8 is selected; The subsequence is classified into species using each tenth reference sequence, and the lowest level species that can be classified is used as the nt verified species identification result of the subsequence, and the lowest value of the similarity and coverage of the second reference sequence used for species classification is used as the similarity and coverage of the subsequence to the species.

7. A nanopore multi-targeted microbial identification system according to claim 6, characterized in that: Also includes: The final identification module is connected to the initial identification module and the nt verification identification module respectively, and is used for final species identification: for each subsequence, there is a corresponding initial species identification result t0 and nt verification species identification result t1. If the similarity and coverage of t1 are not less than t0, t1 is used as the final species identification result of the subsequence. Otherwise, if the similarity of t0 is not less than 97% and the coverage is not less than 80%, t0 is used as the final species identification result of the subsequence. In other cases, it is considered that the subsequence is of low quality and no suitable species classification is filtered out.

8. A nanopore multi-targeted microbial identification system according to any one of claims 5 to 7, characterized in that: Before the data input module and the comparison module, pre-processing is further performed on the sequencing data, including sequencing adapter shearing, low-complexity region splitting, low-quality sequence filtering and / or short sequence filtering. The low-complexity region splitting is specifically: extracting non-low-complexity regions on both sides of the low-complexity region; the short sequence refers to a sequence with a length of less than 80 bp.

9. A computer device, characterized in that: include: Memory for storing computer programs; A processor, used to implement the steps of a nanopore multi-targeted microorganism identification method as described in any one of claims 1 to 4 when executing the computer program.

10. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the steps of a nanopore multi-targeted microorganism identification method as described in any one of claims 1 to 4 are implemented.

Images