DNA storage method for rapidly reading short fragments through nanopore sequencing

Abstract

The invention discloses a DNA (deoxyribonucleic acid) storage method for quickly reading short fragments in nanopore sequencing, which comprises the following steps of: coding an information sequence by using a block code, sparsening a code word fragment, superposing the sparsened code word fragment with a known pseudo-random sequence, combining an equal-length coding bit sequence, mapping the equal-length coding bit sequence into a data DNA sequence, and constructing a label DNA sequence by using the pseudo-random sequence along with a short block error correction code; the method comprises the following steps of: aiming at nanopore sequencing, realizing quick clustering of sequencing reads by utilizing primer recognition and label recognition, constructing a hidden Markov model, deducing the maximum symbol probability by utilizing a forward-backward algorithm, performing upper-layer bit sequence merging on the reads in the clusters, correcting insertion and deletion errors, obtaining a consistent upper-layer bit sequence, and correcting a lower-layer bit sequence in the clusters; and calculating the likelihood probability of the consistency sequence, executing soft decision error correction decoding, and recovering the original data. The method can effectively utilize nanopore sequencing to quickly read short DNA fragments, and has the advantage that a probability-based multi-sequence merging method is used to correct high-insertion deletion errors of nanopore sequencing.

Description

Technical Field

[0001] The present invention relates to the technical field of DNA storage, and in particular to a DNA storage method for rapidly reading out short fragments through nanopore sequencing. Background Art

[0002] As the trend of digitalization, networking and intelligence of information technology continues to deepen, the amount of data has shown explosive growth. Traditional media face many challenges such as cost and energy consumption, and cannot meet the rapidly growing demand for data storage. DNA data storage has ultra-high physical storage density and long-term stability of the medium, and has become a very promising future storage medium. With the rapid development of DNA high-throughput synthesis technology and sequencing technology, it is technically feasible to use DNA sequences to store digital information and recover the original digital information from sequencing reads. In 2021, the Semiconductor Industry Association of the United States released the "Semiconductor 10-Year Plan", and DNA storage became one of the four main media for large-scale data storage in the future. DNA is a biological macromolecule with good stability. Compared with existing magnetic, optical, and electrical storage media, DNA as a data storage medium has the characteristics of small size, high density, and long-lasting medium availability. In terms of density, Columbia University published a paper in Science confirming that the storage density of DNA can reach 215PB / g; Tianjin University used product coding with lower complexity to synthesize more than 270,000 short-chain DNAs, achieving a DNA data storage density of 125PB / g (Science China Life Sciences, 2020). In terms of storage stability, Grass et al. confirmed that if DNA encoding information is encapsulated in materials such as silica, it can be preserved for thousands of years.

[0003] DNA data storage uses modern information coding technology to convert binary digital information into DNA sequences, writes data into physical DNA through high-throughput DNA synthesis, and reads out a large number of DNA molecules in parallel through high-throughput DNA sequencing technology. The storage form of its medium is an oligonucleotide pool, which can be stored in the form of dry powder, solution or other enhanced packaging. When reading data, second-generation high-throughput sequencing such as Illumina and Ion Torrent platforms are generally used, both of which are based on sequencing-by-synthesis technology, and the sequencing process is time-consuming. The third-generation nanopore sequencing can realize real-time single-molecule sequencing. The R10 chip reads out at a rate of 400 bases per second, and can quickly read out short-fragment DNA sequences. In recent years, with the rapid development of solid-state nanopore technology, due to the continuous improvement of the throughput of nanopore sequencing, the use of nanopores to achieve rapid reading of large-scale DNA molecule pools has become a very feasible solution.

[0004] However, nanopore sequencing has a high initial error rate, including: difficult to handle base insertion and deletion errors. Yazdi et al. used a constraint coding and homopolymer check code coding scheme for long DNA sequences to achieve error-free data recovery at 200× sequencing coverage. Tianjin University proposed a watermark code constructed by superimposing LDPC codes on pseudo-random sequences. This method uses a forward-backward algorithm to calculate the offset path for a yeast artificial chromosome with a length of 254kb, identify and correct base insertion and deletion errors, and can recover data at a sequencing coverage of 16.8×. This method is based on the watermark and sparse coding method, which can correct a large number of insertions and deletions. It is a systematic method. The researchers have realized the use of nanopore rapid sequencing to read out data stored in a yeast artificial chromosome with a length of 254kb. Press et al. developed a cascade code coding scheme with an inner code of greedy search decoding hash code (HEDGES), which can also be used to correct insertion and deletion errors, but this method has not been tested in the reagent environment of third-generation nanopore sequencing.

[0005] For nanopore sequencing, researchers proposed a rolling circle concatenation consensus (R2C2) method to generate highly accurate nanopore sequencing consensus sequences. Microsoft researchers used polymerase chain reaction and Gibson assembly technology to assemble short DNA fragments into large DNA fragments. This method achieved data recovery at 22× sequencing coverage, but complex sample preparation reduced the practicality of DNA data storage. However, there is no effective method to directly sequence and read short DNA sequences using nanopore sequencing. Summary of the invention

[0006] The present invention provides a method for storing short DNA fragments that are quickly read out by nanopore sequencing. The present invention performs probability-based multi-sequence merging on sequencing read copies obtained by nanopore sequencing, thereby achieving fast and error-free data recovery, as described below:

[0007] A method for storing DNA for rapid readout of short fragments by nanopore sequencing, the method comprising the following steps:

[0008] (1) The data to be stored is encoded using a block error correction code C1(n1,k1), where k1 is the length of the data to be stored, n1 is the length of the block error correction code codeword, and the codewords are interleaved and segmented. The codewords of some segments are sparsely sequenced and superimposed with a pseudo-noise sequence, and other segments of equal length are combined and mapped into a data DNA sequence according to a preset mapping rule. The label part uses a short block error correction code C2(n2,k2) accompanied by a pseudo-random sequence, where k2 is the bit length before the label part is encoded, and n2 is the length of the short block error correction code codeword. The label range is To distinguish data blocks, primer sequences are added at both ends of the coding DNA sequence;

[0009] (2) The synthesized oligonucleotide pool is subjected to polymerase chain reaction to achieve sample amplification, and the DNA sample is library constructed and sequenced;

[0010] (3) Identify the double-ended primers of the sequencing reads, extract the valid data portion of the reads, and use the watermark sequence w2 of the label part to cyclically shift and identify and correct the insertion and deletion errors in the label sequence. Label identification is used to cluster the reads. Each read in the cluster is demapped into a two-layer bit sequence, where the upper layer bit sequence set is R and the lower layer bit sequence set is S. The data part watermark sequence w1 is used to derive the encoding codeword probability V of the upper layer bit sequence in the cluster, and multiple sequences are merged based on the multi-probability merging method to correct the insertion and deletion errors and obtain a consistent upper layer sequence, which is used to correct the lower layer bit sequence in the cluster to obtain a consistent lower layer sequence. The corrected codeword is sent to the decoder to perform soft decision error correction decoding to restore the original data.

[0011] Wherein, the specific steps of step (1) are as follows:

[0012] (1.1) The block error correction code C1(n1,k1) is used to encode the data to be stored with a length of k1 to generate a codeword sequence with a length of n1. After the codeword sequence with a length of n1 is interleaved, it is divided into p segments of equal length;

[0013] (1.2) The segmented codeword is divided into two layers of bit sequences, the upper layer bit sequence length is 4m, and the lower layer bit sequence length is 5m. The upper layer bit sequence is thinned by converting 4 bits to 5 bits, and then a watermark sequence of equal length is superimposed, and n1 is an integer multiple of 9m;

[0014] (1.3) Map the upper layer bit sequence and the lower layer bit sequence into a coding DNA sequence according to the mapping rules: 00→A, 01→T, 10→G, 11→C;

[0015] (1.4) Using the short block error correction code C2(n2,k2) for the range The label is encoded, k2 is the bit length of the label part before encoding, n2 is the codeword length of the short packet error correction code, and the encoded codeword sequence and the watermark sequence are mapped according to the same mapping rule as step (1.3) to obtain a label DNA sequence with strong error correction capability;

[0016] (1.5) Add the labeled DNA sequence to the left end of the data DNA sequence, and add primer sequences for amplification at both ends.

[0017] The specific steps of step (3) are as follows:

[0018] (3.1) According to the edit distance between the read and the double-ended primer, the sequencing reads are screened, the label part and the data part between the two end primers are intercepted, the label part is demapped to obtain the part corresponding to the known pseudo-random sequence, and the insertion and deletion errors of the label part are identified and corrected by circular shift and dynamic programming, and then error correction decoding is performed to cluster the sequencing reads;

[0019] (3.2) Demap the c read copies in the cluster into two-layer bit sequences, where the upper-layer bit sequence set R = {r 1 ,r 2 ,…,r c}, the lower layer bit sequence set S = {s 1 ,s 2 ,…,s c}, take r from the set R in turn i As the observation vector, combined with the watermark sequence w1, the forward-backward algorithm based on the hidden Markov model is executed to estimate the forward metric and backward metric of each state, as well as the intermediate metric containing the symbol soft information, and output the symbol probability result V = {v1,v2,…,v c}, and further use the multi-sequence merging strategy based on probability merging to obtain the probability v of the consensus symbol within the cluster c , according to the maximum possible probability of each symbol position, the upper layer symbol is inferred to obtain the consensus upper layer bit sequence;

[0020] (3.3) Compare the consensus upper-layer bit sequence after error correction with the set of upper-layer bit sequences obtained from the c read segments in the cluster, identify the error location, correct the insertion and deletion errors of the lower-layer bit sequence set in turn, perform majority voting on the corrected lower-layer bit sequence in the cluster to obtain a consensus lower-layer bit sequence, and concatenate it with the error-corrected upper-layer bit sequence to generate probability information for soft decision decoding, perform error correction decoding, and restore the original data.

[0021] Wherein, the specific steps of step (3.1) are:

[0022] (3.1.1) Align the sequencing reads with the designed double-ended primers, screen the sequencing reads, and determine the boundary positions of the double-ended primers in the sequencing reads;

[0023] (3.1.2) Cut off the label part, use correlation detection to identify the watermark sequence window, and demap the label sequence corresponding to the window to obtain the damaged watermark sequence r 0 and the damaged label bit sequence s 0 ;

[0024] (3.1.3) The damaged watermark sequence r 0The error-free watermark sequence w2 is cyclically shifted in sequence, and the insertion and deletion positions are identified using dynamic programming to correct the damaged label bit sequence s 0 , the corrected results are subjected to error correction decoding, and the read segments are clustered accordingly.

[0025] Wherein, the specific steps of step (3.2) are:

[0026] (3.2.1) According to the error characteristics of nanopore sequencing, estimate the insertion error probability P and deletion error probability of sequencing reads

[0027] i P d , and the substitution error probability P s Construct an error transmission model, and read the sequence r from the upper layer i , i∈[1,c], as the observation vector and the base offset as the hidden state, and perform the forward-backward algorithm based on the hidden Markov model to estimate the forward metric and backward metric of each state, as well as the intermediate metric containing the soft information of the symbol;

[0028] (3.2.2) For a DNA sequence of length N, the number of symbols is N / 5. Using the soft-decision forward-backward algorithm, the symbol probability distribution v = {p1, p2, …, p N / 5}, where p j is the probability distribution of the jth symbol, p j =(p j,0 ,p j,1 ...,p j,k ),

[0029] (3.2.3) According to the probability information V calculated in each cluster, V = {v1, v2, ..., v c}, the corresponding positions are multiplied and normalized in sequence, and the consensus symbol probability v after each cluster is merged is output c ={p1′,p2′,…,p N / 5 ′}, where p j ′ is the probability distribution of the jth symbol, p j ′=(p j,0 ′,p j,1 ′...,p j,k ′),

[0030] (3.2.4) According to the maximum possible probability max(p j,0 ′,p j,1 ′,…,p j,15 ′), infer the upper-layer symbol sequence and obtain the consensus upper-layer bit sequence.

[0031] Wherein, the specific steps of step (3.3) are:

[0032] (3.3.1) Compare the consensus upper bit sequence after error correction with the original upper bit sequence carrying errors demapped from the c sequencing reads in the cluster to identify the location where the insertion and deletion errors occurred in each read;

[0033] (3.3.2) According to the identified insertion and deletion error positions, assist in correcting the damaged codeword sequence of the corresponding lower layer, perform majority voting on all corrected lower layer bit sequences in the cluster, and calculate the consensus lower layer bit sequence;

[0034] (3.3.3) combining the consensus upper layer bit sequence and the consensus lower layer bit sequence, and calculating the probabilistic soft information for soft decision decoding;

[0035] (3.3.4) The probabilistic soft information is sent to the decoder, and the logarithmic domain belief propagation code algorithm is executed to correct the remaining substitution errors, obtain the decoded codeword, and restore the original data.

[0036] The beneficial effects of the technical solution provided by the present invention are:

[0037] 1. The present invention uses a coding strategy with high coding efficiency in the data part and strong error correction capability in the label part to encode the data file into a DNA sequence; a synthetic oligonucleotide pool is used as a data storage medium, and sequencing reads are quickly obtained using nanopore sequencing; and rapid clustering of sequencing reads is achieved using double-ended primer recognition and highly reliable label recognition;

[0038] 2. The present invention aims at the problem of high insertion and deletion errors, and uses the symbol probability calculated by the soft decision forward-backward algorithm to realize probability-based multi-sequence merging, so that the error correction algorithm can play a more accurate role;

[0039] 3. The present invention provides a solution for oligonucleotide pool storage using nanopore sequencing technology to achieve rapid readout. BRIEF DESCRIPTION OF THE DRAWINGS

[0040] Figure 1 A block diagram of a DNA storage method for rapid readout of short fragments by nanopore sequencing provided by the present invention;

[0041] Figure 2 A flow chart for designing a data DNA sequence with high coding efficiency of the present invention;

[0042] Figure 3 Design flow chart for the labeled DNA sequence with strong error correction capability of the present invention;

[0043] Figure 4 The data reading flow chart of the present invention is as follows;

[0044] Figure 5 A flowchart for performing double-ended primer identification and high-reliability label identification in the present invention;

[0045] Figure 6 This is a multi-sequence merging flow chart based on probability merging of the present invention;

[0046] Figure 7 A flowchart of the present invention assisting lower layer bit error correction;

[0047] Figure 8 A length distribution diagram of the sequencing reads provided by the present invention after the effective data portion is cut off;

[0048] Fig. 9 A distribution diagram of the number of copies within a cluster after the read segment is clustered after label identification provided by the present invention;

[0049] Fig.10 An initial error distribution diagram of sequencing reads provided by the present invention;

[0050] Fig.11 This is a bit error rate distribution diagram after multiple sequences are combined based on probability combination provided by the present invention. DETAILED DESCRIPTION

[0051] In order to make the objectives, technical solutions and advantages of the present invention more clear, the embodiments of the present invention are described in further detail below.

[0052] In view of the above problems, an embodiment of the present invention proposes a coding strategy of a watermark code constructed by superimposing an error correction code on a single layer of pseudo-random sequence, and the coding DNA sequence is synthesized into a short oligonucleotide as a storage medium. After linking the library and nanopore signal recognition, a sequencing read with a large number of base insertion and deletion errors is obtained. When the data is read out, combined with the multi-copy characteristics of the coding DNA sequence, a method is proposed to use watermarks to derive the probability of coding codewords, which is used to realize multi-sequence merging in a multi-probability merging manner. Compared with the direct use of the MSA algorithm, it has the advantages of better matching the coding algorithm and stronger error correction capabilities, and supports the storage mode of short-fragment DNA data that is quickly read out by nanopore sequencing.

[0053] The embodiment of the present invention proposes an encoding and recovery method for storing information using short-fragment DNA. A coding strategy in which the data part has high coding efficiency and the label part has strong error correction capability is used to encode the data file into a data DNA sequence. A synthetic oligonucleotide pool is used as a data storage medium, and sequencing reads are quickly obtained using nanopore sequencing. The reads are clustered through double-ended primer recognition and highly reliable label recognition, and the symbol probability calculated by the soft-decision forward-backward algorithm is used to achieve probability-based multi-sequence merging, correct insertion and deletion errors in nanopore sequencing reads, and symbol likelihood information also improves error correction decoding performance to achieve error-free recovery.

[0054] The implementation of the embodiment of the present invention is described in detail below with reference to the accompanying drawings.

[0055] This embodiment describes in detail a method for storing short DNA fragments quickly read by nanopore sequencing proposed in an embodiment of the present invention. Figure 1 , specifically including the following steps:

[0056] (1) The data to be stored is encoded using a block error correction code C1(n1,k1), where k1 is the length of the data to be stored, n1 is the length of the block error correction code codeword, and the codewords are interleaved and divided into p segments of equal length. The segmented codeword sequence is superimposed with a pseudo-noise sequence and then mapped into a data DNA sequence according to a specific mapping rule. The label part uses a short block error correction code C2(n2,k2) accompanied by a pseudo-random sequence, where k2 is the bit length before the label part is encoded, and n2 is the length of the short block error correction code codeword. The label range is To distinguish data blocks, primer sequences are added at both ends of the coding DNA sequence;

[0057] (2) The synthesized oligonucleotide pool is subjected to multiple rounds of PCR thermal cycling reactions to amplify the target DNA fragments, and further, the DNA fragments are subjected to library construction and third-generation sequencing;

[0058] (3) Identify the double-ended primers of the sequencing reads, extract the valid data portion of the reads, and use the watermark sequence w2 of the label part to cyclically shift and identify and correct the insertion and deletion errors in the label sequence, so as to achieve highly reliable label recognition for read clustering. Each read in the cluster is demapped into a two-layer bit sequence, where the upper layer bit sequence set is R and the lower layer bit sequence set is S. The data part watermark sequence w1 is used to derive the encoding codeword probability V of the upper layer bit sequence in the cluster, and multiple sequences are merged based on the multi-probability merging method to correct the insertion and deletion errors and obtain a consistent upper layer sequence, which is used to correct the lower layer bit sequence in the cluster to obtain a consistent lower layer sequence. The corrected codeword is sent to the decoder to perform soft decision error correction decoding to restore the original data.

[0059] Among them, Figure 2 and Figure 3 As shown, the data to be stored is encoded using the packet error correction code C1 (n1, k1), k1 is the length of the data to be stored, n1 is the length of the packet error correction code codeword, and the codewords are interleaved and divided into p segments of equal length. After the segmented codeword sequence is superimposed with the pseudo-noise sequence, it is mapped into a data DNA sequence according to a specific mapping rule. The label part uses a short packet error correction code C2 (n2, k2) accompanied by a pseudo-random sequence, k2 is the bit length before the label part is encoded, n2 is the length of the short packet error correction code codeword, and the label range is To distinguish data blocks, primer sequences are added at both ends of the coding DNA sequence. The specific steps are as follows:

[0060] (1.1) The block error correction code C1(n1,k1) is used to encode the data to be stored with a length of k1 to generate a codeword sequence with a length of n1. After the codeword sequence with a length of n1 is interleaved, it is divided into p segments of equal length;

[0061] (1.2) The segmented codeword is divided into two layers of bit sequences according to the proportional coefficient m. The length of the upper layer bit sequence is 4m, and the length of the lower layer bit sequence is 5m. The upper layer bit sequence is sparsely processed by converting 4 bits to 5 bits, and then a watermark sequence of equal length is superimposed;

[0062] (1.3) Map the upper layer bit sequence and the lower layer bit sequence into a coding DNA sequence according to the mapping rules: 00→A, 01→T, 10→G, 11→C;

[0063] (1.4) Using the short block error correction code C2(n2,k2) for the range The label is encoded, k2 is the bit length of the label part before encoding, n2 is the codeword length of the short packet error correction code, and the encoded codeword sequence and the watermark sequence are mapped according to the same mapping rule as step (1.3) to obtain a label DNA sequence with strong error correction capability;

[0064] (1.5) Add the labeled DNA sequence to the left end of the data DNA sequence, and add primer sequences for amplification at both ends.

[0065] Among them, Figure 4 As shown, the double-ended primers of the sequencing reads are identified, the valid data part of the reads is intercepted, and the watermark sequence w2 of the label part is used for cyclic shift identification and correction of insertion and deletion errors in the label sequence, so as to realize highly reliable label recognition for read clustering, and each read in the cluster is demapped into a double-layer bit sequence, wherein the upper layer bit sequence set is R and the lower layer bit sequence set is S. The watermark sequence w1 of the data part is used to derive the encoding codeword probability V of the upper layer bit sequence in the cluster, and multiple sequences are merged based on the multi-probability merging method, and the insertion and deletion errors are corrected to obtain a consistent upper layer sequence, which is used to correct the lower layer bit sequence in the cluster to obtain a consistent lower layer sequence, and the corrected codeword is sent to the decoder to perform soft decision error correction decoding to restore the original data. The specific steps are as follows:

[0066] (3.1) According to the edit distance between the read and the double-ended primer, the sequencing reads are screened, the label part and the data part between the two end primers are intercepted, the label part is demapped to obtain the part corresponding to the known pseudo-random sequence, and the insertion and deletion errors of the label part are identified and corrected by circular shift and dynamic programming, and then error correction decoding is performed to cluster the sequencing reads;

[0067] (3.2) Demap the c read copies in the cluster into two-layer bit sequences, where the upper-layer bit sequence set R = {r 1 ,r 2 ,…,r c}, the lower layer bit sequence set S = {s 1 ,s 2 ,…,s c}, take r from the set R in turn i As the observation vector, combined with the watermark sequence w1, the forward-backward algorithm based on the hidden Markov model is executed to estimate the forward metric and backward metric of each state, as well as the intermediate metric containing the symbol soft information, and output the symbol probability result V = {v1,v2,…,v c}, and further use the multi-sequence merging strategy based on probability merging to obtain the probability v of the consensus symbol within the cluster c , according to the maximum possible probability of each symbol position, the upper layer symbol is inferred to obtain the consensus upper layer bit sequence;

[0068] (3.3) Compare the consensus upper-layer bit sequence after error correction with the set of upper-layer bit sequences obtained from the c read segments in the cluster, identify the error location, correct the insertion and deletion errors of the lower-layer bit sequence set in turn, perform majority voting on the corrected lower-layer bit sequence in the cluster to obtain a consensus lower-layer bit sequence, and concatenate it with the error-corrected upper-layer bit sequence to generate probability information for soft decision decoding, perform error correction decoding, and restore the original data.

[0069] Among them, Figure 5 As shown in the figure, according to the edit distance between the read segment and the double-ended primer, the sequencing read segment is screened, the label part and the data part between the two end primers are intercepted, the label part is demapped to obtain the part corresponding to the known pseudo-random sequence, and the insertion and deletion errors of the label part are identified and corrected by circular shift and dynamic programming, and then error correction decoding is performed to cluster the sequencing read segment. The specific steps are as follows:

[0070] (3.1.1) Align the sequencing reads with the designed double-ended primers, screen the sequencing reads, and determine the boundary positions of the double-ended primers in the sequencing reads;

[0071] (3.1.2) Cut off the label part, use correlation detection to identify the watermark sequence window, and demap the label sequence corresponding to the window to obtain the damaged watermark sequence r 0 and the damaged label bit sequence s 0 ;

[0072] (3.1.3) The damaged watermark sequence r 0 The error-free watermark sequence w2 is cyclically shifted in sequence, and the insertion and deletion positions are identified using dynamic programming to correct the damaged label bit sequence s 0, the corrected results are subjected to error correction decoding, and the read segments are clustered accordingly.

[0073] Among them, Figure 6 As shown, the c read segment copies in the cluster are demapped into two-layer bit sequences, where the upper layer bit sequence set R = {r 1 ,r 2 ,…,r c}, the lower layer bit sequence set S = {s 1 ,s 2 ,…,s c}, take r from the set R in turn i As the observation vector, combined with the watermark sequence w1, the forward-backward algorithm based on the hidden Markov model is executed to estimate the forward metric and backward metric of each state, as well as the intermediate metric containing the symbol soft information, and output the symbol probability result V = {v1,v2,…,v c}, and further use the multi-sequence merging strategy based on probability merging to obtain the probability v of the consensus symbol within the cluster c , based on the maximum possible probability of each symbol position, infer the upper-layer symbol and obtain the consensus upper-layer bit sequence. The specific steps are:

[0074] (3.2.1) According to the error characteristics of nanopore sequencing, estimate the insertion error probability P of the sequencing read segment i , deletion error probability P d , and the substitution error probability P s Construct an error transmission model, and read the sequence r from the upper layer i , i∈[1,c], as the observation vector and the base offset as the hidden state, and perform the forward-backward algorithm based on the hidden Markov model to estimate the forward metric and backward metric of each state, as well as the intermediate metric containing the soft information of the symbol;

[0075] (3.2.2) For a DNA sequence of length N, the number of symbols is N / 5. Using the soft-decision forward-backward algorithm, the symbol probability distribution v = {p1, p2, …, p N / 5}, where p j is the probability distribution of the jth symbol, p j =(p j,0 ,p j,1 ...,p j,k ),

[0076] (3.2.3) According to the probability information V calculated in each cluster, V = {v1, v2, ..., v c}, the corresponding positions are multiplied and normalized in sequence, and the consensus symbol probability v after each cluster is merged is output c ={p1′,p2′,…,pN / 5 ′}, where p j ′ is the probability distribution of the jth symbol, p j ′=(p j,0 ′,p j,1 ′...,p j,k ′),

[0077] (3.2.4) According to the maximum possible probability max(p j,0 ′,p j,1 ′,…,p j,15 ′), infer the upper-layer symbol sequence and obtain the consensus upper-layer bit sequence.

[0078] Among them, Figure 7 As shown in the figure, the consensus upper-layer bit sequence after error correction is compared with the upper-layer bit sequence set obtained from the c read segments in the cluster, the error position is identified, the insertion and deletion errors of the lower-layer bit sequence set are corrected in turn, the corrected lower-layer bit sequence in the cluster is subjected to majority voting to obtain the consensus lower-layer bit sequence, which is then concatenated with the error-corrected upper-layer bit sequence to generate probability information of soft decision decoding, perform error correction decoding, and restore the original data. The specific steps are as follows:

[0079] (3.3.1) Compare the consensus upper bit sequence after error correction with the original upper bit sequence carrying errors demapped from the c sequencing reads in the cluster to identify the location where the insertion and deletion errors occurred in each read;

[0080] (3.3.2) According to the identified insertion and deletion error positions, assist in correcting the damaged codeword sequence of the corresponding lower layer, perform majority voting on all corrected lower layer bit sequences in the cluster, and calculate the consensus lower layer bit sequence;

[0081] (3.3.3) combining the consensus upper layer bit sequence and the consensus lower layer bit sequence, and calculating the probabilistic soft information for soft decision decoding;

[0082] (3.3.4) The probabilistic soft information is sent to the decoder, and the logarithmic domain belief propagation code algorithm is executed to correct the remaining substitution errors, obtain the decoded codeword, and restore the original data. Specific embodiments

[0084] Specific examples are given below to illustrate the feasibility of a method for storing short-fragment DNA data that is quickly read out by nanopore sequencing provided by the present invention.

[0085] The embodiment of the present invention uses a coding method of cascading LDPC codes and watermark codes to encode six audio segments of 1,320,907 bytes, 1,320,646 bytes, 1,321,986 bytes, 1,320,237 bytes, 1,320,010 bytes and 1,320,580 bytes into DNA sequences and synthesize them into oligonucleotide pools. The oligonucleotide pools are named pool1, pool2, pool3, pool4, pool5 and pool6. The code length of the LDPC code is 64800 bits, the information bit length is 54000 bits, the code rate is 5 / 6, the code rate of the watermark code is 4 / 5, and the overall efficiency is 4 / 3 bits per base.

[0086] The encoding method of LDPC code and watermark code concatenation is as follows:

[0087] (1.1) The 54,000-bit information sequence is encoded with an LDPC code. The length of the encoded codeword sequence is 64,800 bits. Each audio segment is encoded into 196 codewords. Each codeword is segmented into a fixed length of 423 bits after interleaving. Each codeword is divided into 153 segments.

[0088] (1.2) For each bit sequence, the first 188 bits are taken as the upper-layer bit sequence, and the last 235 bits are taken as the lower-layer bit sequence. After the upper-layer bit sequence is sparsely converted from 4 bits to 5 bits, it is superimposed with the watermark sequence to obtain a 235-bit upper-layer bit sequence with superimposed watermark. Then, one bit is selected from the lower-layer bit sequence in turn to form a bit pair.

[0089] (1.3) The bit pairs are mapped to bases according to the mapping rules of 00→A, 01→T, 10→G, 11→C, and a data DNA sequence with a length of 235 bases is obtained. The process of superimposing the watermark of the first segment of 423 bits of the first codeword and mapping it to a base sequence is shown in Table 1;

[0090] (1.4) Using the BCH code (31, 21) coding strategy, 1-2 15 The binary label sequence of the range is encoded, and the encoded codeword is combined with the watermark sequence of equal length. After mapping, the label DNA sequence is obtained and added to the left end of the data DNA sequence. The primer sequence with a length of 20 bases is added to the left end of the label DNA sequence and the right end of the data DNA sequence respectively, and finally the coding DNA sequence with a length of 300 bases is obtained.

[0091] Table 1 The process of superimposing watermarks on codeword sequences and mapping them into base sequences

[0092]

[0093]

[0094] In the embodiment of the present invention, the DNA sequence is synthesized into an oligonucleotide pool as a storage medium, and after library construction and third-generation nanopore sequencing, nanopore sequencing reads are obtained. The method for recovering data from nanopore sequencing reads in the embodiment of the present invention is specifically as follows:

[0095] (2.1) The start and end positions of the payload sequence are determined by comparing the edit distance between the double-end product and the sequencing read segment, and the sequencing read segment is truncated according to the start and end positions. In the embodiment of the present invention, the length distribution of the data after truncation is as follows: Figure 8 As shown, it can be seen that the read length is around 260 bases;

[0096] (2.2) The label part is intercepted and demapped to obtain the damaged label bit sequence and damaged watermark sequence. The insertion and deletion errors of the label part are identified and corrected by cyclic shift and dynamic programming, and the error correction decoding of the label part is performed. The sequencing reads with the same label are clustered according to the identification results. The embodiment of the present invention performs label identification on the sequencing reads corresponding to the 6 oligonucleotide pools respectively, and the identification results are shown in Table 2. The 5x sequencing data is used, and the coverage distribution of the reads in the cluster after label identification is shown in Fig. 9 As shown, it can be seen that the average coverage of the reads in the cluster is 4.5x, and the highest coverage reaches 21x. Among them, the two reads corresponding to the coding DNA sequence in Table 1 after label recognition are shown in Table 3, the length of read 1 is 235 bases, and the length of read 2 is 236 bases;

[0097] Table 2. Proportion of successful label recognition

[0098] 3x 4x 5x 6x pool1 87.76％ 87.76％ 87.81％ 87.82％ pool2 89.13％ 89.09％ 89.34％ 89.22％ pool3 88.14％ 88.22％ 88.16％ 88.14％ pool4 89.17％ 89.29％ 89.31％ 89.34％ pool5 88.88％ 88.94％ 88.90％ 88.90％ pool6 90.16％ 90.04％ 90.2％ 90.14％

[0099] Table 3 Two reads in a cluster after label identification

[0100]

[0101]

[0102] (2.3) In each cluster, the upper layer bit sequence set R takes r i As the observation vector, combined with the watermark sequence w1, the forward-backward algorithm based on the hidden Markov model is executed to estimate the forward metric and backward metric of each state, as well as the intermediate metric containing the symbol soft information. The probability of the consensus symbol in the cluster is obtained by using the multi-sequence merging strategy based on probability merging. According to the maximum possible probability of each symbol position, the upper-level symbol is inferred to obtain the consensus upper-level bit sequence;

[0103] Fig.10The initial error probability distribution of DNA sequences of the six oligonucleotide pools is shown. It can be seen that the total base error rate of each oligonucleotide pool is basically the same and all below 1.4%. Based on the error characteristics of the above sequencing reads, a hidden Markov model is established, and the insertion error probability P is set. i =0.004, deletion error probability P d =0.005, substitution error probability P s =0.006, maximum offset X max = 10, the maximum number of consecutive insertions I = 2. The symbol likelihood probability is estimated using the soft decision forward-backward algorithm, the upper symbol probability of each cluster is merged, the consensus and normalized symbol probability are obtained, and the upper symbol is inferred according to the maximum possible probability of each symbol position to obtain the consensus upper bit sequence, where the symbol likelihood probability is calculated for the two read segments shown in Table 3, and the process of multi-sequence merging based on probability merging is shown in Table 4;

[0104] Table 4 Symbol probability calculation and merging of read segments in the same cluster

[0105]

[0106]

[0107]

[0108] (2.4) For the lower-layer bits of the read segment, compare the upper-layer bit sequences before and after error correction, identify the error position of the read segment, and correct the insertion and deletion errors of the lower-layer bits at the corresponding position. Perform majority voting on the corrected lower-layer bit sequence in the cluster to obtain a consistent lower-layer bit sequence. The error correction process for the lower-layer bits of the read segment in Table 3 is shown in Table 5. The lower-layer bit sequence after error correction is concatenated with the upper-layer bit sequence after error correction to generate probability information for soft decision decoding, perform error correction decoding, and restore the original data.

[0109] Fig.11 The comparison of the upper and lower bit error rates of the sequencing reads of the six oligonucleotide pools in the embodiment of the present invention after error correction is shown. It can be seen that the upper and lower bit error rates of the six groups of sequencing reads after error correction are basically the same, all lower than the 1.5% initial error rate preset in the embodiment of the present invention, which fully verifies that the probability-based multi-sequence merging method adopted can effectively deal with the problem of high error rate of nanopore sequencing.

[0110] Table 5 Lower layer bit error correction process

[0111]

[0112]

[0113] The embodiment of the present invention compares the proposed error correction method for correcting upper layer bit errors and assisting in correcting lower layer bit errors with the multi-sequence merging method. The multi-sequence merging method first uses MAFFT to perform multiple sequence alignment, and then obtains the merged DNA sequence through majority voting. The merged DNA sequence uses the forward and backward algorithm to further correct insertion and deletion errors, and is input into the decoder for decoding.

[0114] The bit error rates after error correction by the two methods are shown in Table 6. It can be seen that for the data of the six oligonucleotide pools, at 10x coverage, the proposed error correction method achieves a lower error rate than the method based on multi-sequence merging. The running time of the error correction of the two methods is shown in Table 7. The method based on multi-sequence merging is faster than the proposed error correction method. In the embodiment of the present invention, each oligonucleotide pool corresponds to 196 codewords. The codeword decoding recovery after error correction by the two methods is shown in Table 8, which shows the number of codewords that can be correctly decoded in each oligonucleotide pool under different sequencing coverages. It can be seen that the method based on multi-sequence merging relies on high sequencing coverage, and the error correction method based on probability merging proposed in the embodiment of the present invention is better than the method based on multi-sequence merging in error correction effect at low sequencing coverage.

[0115] The embodiment of the present invention corrects errors in short DNA sequences sequenced in the literature of DNA information storage solutions based on nanopore sequencing to verify the feasibility of the present invention.

[0116] Table 6 Comparison of error rates between this method and the method based on multiple sequence merging at 10x coverage

[0117]

[0118] Table 7 Comparison of running time between this method and the method based on multiple sequence merging at 10x coverage (unit: seconds)

[0119]

[0120] Table 8 Comparison of the number of codewords successfully decoded by this method and the method based on multi-sequence merging

[0121]

[0122] Those skilled in the art will appreciate that the accompanying drawing is only a schematic diagram of a preferred embodiment, and the serial numbers of the embodiments of the present invention are only for description and do not represent the advantages or disadvantages of the embodiments.

[0123] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principle of the present invention should be included in the protection scope of the present invention.

Claims

1. A method for storing DNA for rapid readout of short fragments by nanopore sequencing, characterized in that: The method comprises the following steps: (1) The data to be stored is encoded using a block error correction code C1(n1,k1), where k1 is the length of the data to be stored, n1 is the length of the block error correction code codeword, and the codewords are interleaved and segmented. The codewords of some segments are sparsely sequenced and superimposed with a pseudo-noise sequence, and other segments of equal length are combined and mapped into a data DNA sequence according to a preset mapping rule. The label part uses a short block error correction code C2(n2,k2) accompanied by a pseudo-random sequence, where k2 is the bit length before the label part is encoded, and n2 is the length of the short block error correction code codeword. The label range is To distinguish data blocks, primer sequences are added at both ends of the coding DNA sequence; (2) The synthesized oligonucleotide pool is subjected to polymerase chain reaction to achieve sample amplification, and the DNA sample is library constructed and sequenced; (3) Identify the double-ended primers of the sequencing reads, extract the valid data portion of the reads, and use the watermark sequence w2 of the label part to cyclically shift and identify and correct the insertion and deletion errors in the label sequence. Label identification is used to cluster the reads. Each read in the cluster is demapped into a two-layer bit sequence, where the upper layer bit sequence set is R and the lower layer bit sequence set is S. The data part watermark sequence w1 is used to derive the encoding codeword probability V of the upper layer bit sequence in the cluster, and multiple sequences are merged based on the multi-probability merging method to correct the insertion and deletion errors and obtain a consistent upper layer sequence, which is used to correct the lower layer bit sequence in the cluster to obtain a consistent lower layer sequence. The corrected codeword is sent to the decoder to perform soft decision error correction decoding to restore the original data.

2. A method for storing DNA for rapid readout of short fragments by nanopore sequencing according to claim 1, characterized in that: The specific steps of step (1) are as follows: (1.1) The block error correction code C1(n1,k1) is used to encode the data to be stored with a length of k1 to generate a codeword sequence with a length of n1. After the codeword sequence with a length of n1 is interleaved, it is divided into p segments of equal length; (1.2) The segmented codeword is divided into two layers of bit sequences, the upper layer bit sequence length is 4m, and the lower layer bit sequence length is 5m. The upper layer bit sequence is thinned by converting 4 bits to 5 bits, and then a watermark sequence of equal length is superimposed, and n1 is an integer multiple of 9m; (1.3) Map the upper layer bit sequence and the lower layer bit sequence into a coding DNA sequence according to the mapping rules: 00→A, 01→T, 10→G, 11→C; (1.4) Using the short block error correction code C2(n2,k2) for the range The label is encoded, k2 is the bit length of the label part before encoding, n2 is the codeword length of the short packet error correction code, and the encoded codeword sequence and the watermark sequence are mapped according to the same mapping rule as step (1.3) to obtain a label DNA sequence with strong error correction capability; (1.5) Add the labeled DNA sequence to the left end of the data DNA sequence, and add primer sequences for amplification at both ends.

3. A method for storing DNA for rapid readout of short fragments by nanopore sequencing according to claim 1, characterized in that: The specific steps of step (3) are: (3.1) According to the edit distance between the read and the double-ended primer, the sequencing reads are screened, the label part and the data part between the two end primers are intercepted, the label part is demapped to obtain the part corresponding to the known pseudo-random sequence, and the insertion and deletion errors of the label part are identified and corrected by circular shift and dynamic programming, and then error correction decoding is performed to cluster the sequencing reads; (3.2) Demap the c read copies in the cluster into two-layer bit sequences, where the upper-layer bit sequence set R = {r 1 ,r 2 ,…,r c }, the lower layer bit sequence set S = {s 1 ,s 2 ,…,s c }, take r from the set R in turn i As the observation vector, combined with the watermark sequence w1, the forward-backward algorithm based on the hidden Markov model is executed to estimate the forward metric and backward metric of each state, as well as the intermediate metric containing the symbol soft information, and output the symbol probability result V = {v1,v2,…,v c }, and further use the multi-sequence merging strategy based on probability merging to obtain the probability v of the consensus symbol within the cluster c , according to the maximum possible probability of each symbol position, the upper layer symbol is inferred to obtain the consensus upper layer bit sequence; (3.3) Compare the consensus upper-layer bit sequence after error correction with the set of upper-layer bit sequences obtained from the c read segments in the cluster, identify the error location, correct the insertion and deletion errors of the lower-layer bit sequence set in turn, perform majority voting on the corrected lower-layer bit sequence in the cluster to obtain a consensus lower-layer bit sequence, and concatenate it with the error-corrected upper-layer bit sequence to generate probability information for soft decision decoding, perform error correction decoding, and restore the original data.

4. A method for storing DNA for rapid readout of short fragments by nanopore sequencing according to claim 3, characterized in that: The specific steps of step (3.1) are: (3.1.1) Align the sequencing reads with the designed double-ended primers, screen the sequencing reads, and determine the boundary positions of the double-ended primers in the sequencing reads; (3.1.2) Cut off the label part, use correlation detection to identify the watermark sequence window, and demap the label sequence corresponding to the window to obtain the damaged watermark sequence r 0 and the damaged label bit sequence s 0 ; (3.1.3) The damaged watermark sequence r 0 The error-free watermark sequence w2 is cyclically shifted in sequence, and the insertion and deletion positions are identified using dynamic programming to correct the damaged label bit sequence s 0 , the corrected results are subjected to error correction decoding, and the read segments are clustered accordingly.

5. A method for storing DNA for rapid readout of short fragments by nanopore sequencing according to claim 3, characterized in that: The specific steps of step (3.2) are: (3.2.1) According to the error characteristics of nanopore sequencing, estimate the insertion error probability P and deletion error probability i of the sequencing read segment P d , and the substitution error probability P s Construct an error transmission model, and read the sequence r from the upper layer i , i∈[1,c], as the observation vector and the base offset as the hidden state, and perform the forward-backward algorithm based on the hidden Markov model to estimate the forward metric and backward metric of each state, as well as the intermediate metric containing the soft information of the symbol; (3.2.2) For a DNA sequence of length N, the number of symbols is N / 5. Using the soft-decision forward-backward algorithm, the symbol probability distribution v = {p1, p2, …, p N / 5 }, where p j is the probability distribution of the jth symbol, p j =(p j,0 ,p j,1 ...,p j,k ), (3.2.3) According to the probability information V calculated in each cluster, V = {v1, v2, ..., v c }, the corresponding positions are multiplied and normalized in sequence, and the consensus symbol probability v after each cluster is merged is output c ={p1′,p2′,…,p N / 5 ′}, where p j ′ is the probability distribution of the jth symbol, p j ′=(p j,0 ′,p j,1 ′...,p j,k ′), (3.2.4) According to the maximum possible probability max(p j,0 ′,p j,1 ′,...,p j,15 ′), infer the upper-layer symbol sequence and obtain the consensus upper-layer bit sequence.

6. A method for storing DNA for rapid readout of short fragments by nanopore sequencing according to claim 3, characterized in that: The specific steps of step (3.3) are: (3.3.1) Compare the consensus upper bit sequence after error correction with the original upper bit sequence carrying errors demapped from the c sequencing reads in the cluster to identify the location where the insertion and deletion errors occurred in each read; (3.3.2) According to the identified insertion and deletion error positions, assist in correcting the damaged codeword sequence of the corresponding lower layer, perform majority voting on all corrected lower layer bit sequences in the cluster, and calculate the consensus lower layer bit sequence; (3.3.3) combining the consensus upper layer bit sequence and the consensus lower layer bit sequence, and calculating the probabilistic soft information for soft decision decoding; (3.3.4) The probabilistic soft information is sent to the decoder, and the logarithmic domain belief propagation code algorithm is executed to correct the remaining substitution errors, obtain the decoded codeword, and restore the original data.

Images