Filtering system and method for searching genetic data based on memory

Abstract

A screening system and method based on memory search gene data is disclosed. The screening system includes a coding module, a screening module and an extraction module. By performing pre-alignment screening on the read sequences, the read sequences related to the target reference sequence are accurately extracted, which effectively reduces the amount of data transmission and improves the overall computing efficiency.

Claims

1. A screening system for memory-based gene search data, the screening system comprising: a coding module configured to receive and encode multiple read sequences to generate multiple coded data, wherein the coded data is divided into multiple data groups, each data group including a read sequence index corresponding to a bit address of a corresponding read sequence stored in a storage device; a screening module coupled to the coding module, the screening module configured to compare the coded data with reference data to screen out at least one relevant coded data; and an extraction module coupled to the screening module, the extraction module configured to extract a relevant read sequence stored in the storage device according to the read sequence index corresponding to the at least one relevant coded data; wherein in the screening module, comparing the reference data with the coded data to screen out the at least one relevant coded data includes: Each coded data is subjected to a bitwise logical operation with the reference data, and the results of the bitwise logical operation are summed to obtain a similarity score; the similarity score is compared with a threshold. If the similarity score is lower than the threshold, it is determined that the coded data corresponding to the similarity score is not similar enough to the reference data. If the similarity score is greater than or equal to the threshold, it is determined that the coded data corresponding to the similarity score is related to the reference data, so as to filter out at least one related coded data.

2. The screening system based on memory search gene data as described in claim 1, wherein the screening system further includes a read mapping module coupled to the memory, and the read mapping module is configured to align the relevant read sequence with a reference sequence to produce an alignment result, wherein the reference sequence encodes the reference data.

3. The screening system based on memory search gene data as described in claim 2, wherein the screening system further includes a variant detection module coupled to the read mapping module, and the variant detection module is configured to determine the difference between the relevant read sequence and the reference sequence based on the alignment result, so as to generate a variant detection result.

4. The memory-based gene data screening system as described in claim 1, wherein the encoding module is configured to encode the read sequences having a text format into encoded data having a bit vector format.

5. The memory-based gene data screening system as described in claim 1, wherein the threshold is configured with a fault-tolerant mechanism, and the equation for the threshold is: where is the threshold, is the minimum bit vector similarity score that the encoded data that perfectly matches the reference data can achieve, is the amount of variation allowed between the encoded data and the reference data, and is the DNA fragment size.

6. The screening system based on memory-based gene data search as described in claim 1, wherein the screening module is integrated into a non-volatile memory.

7. The screening system based on memory search gene data as described in claim 1, wherein the screening system further includes an allocation module coupled to the coding module and the storage, and the allocation module is configured to generate a plurality of read sequences corresponding to at least one DNA fragment and transmit the read sequences to the storage and the coding module.

8. The screening system based on memory search gene data as described in claim 7, wherein the screening system further includes a sequencing module coupled to the allocation module, and the sequencing module is configured to process a sample to generate the at least one DNA fragment of the sample, and the screening system is configured to synchronously execute the generation of the read sequences by the allocation module and the generation of the encoded data by the encoding module.

9. A method for screening gene data based on memory search, the method comprising: receiving and encoding multiple read sequences through an encoding module to generate multiple encoded data, wherein the encoded data is divided into multiple data groups, each data group including a read sequence index, the read sequence index corresponding to a bit address of a corresponding read sequence stored in a memory; using a screening module to compare the encoded data with reference data to screen out at least one relevant encoded data; and using an extraction module to extract a corresponding relevant read sequence stored in the memory based on the read sequence index corresponding to the at least one relevant encoded data; wherein using the screening module to compare the encoded data with the reference data to screen out the at least one relevant encoded data includes: Each coded data is subjected to a bitwise logical operation with the reference data, and the results of the bitwise logical operation are summed to obtain a similarity score; the similarity score is compared with a threshold. If the similarity score is lower than the threshold, it is determined that the coded data corresponding to the similarity score is not similar enough to the reference data. If the similarity score is greater than or equal to the threshold, it is determined that the coded data corresponding to the similarity score is related to the reference data, so as to filter out at least one related coded data.

10. The method for screening gene data based on memory search as described in claim 9, wherein after the extraction module extracts a related read sequence stored in the memory according to the read sequence index corresponding to the at least one related encoded data, the screening method further includes: aligning the related read sequence with a reference sequence through a read mapping module to generate an alignment result, wherein the reference sequence encodes the reference data.

11. The screening method for memory search gene data as described in claim 10, wherein after the read mapping module aligns the relevant read sequence with the reference sequence to generate the alignment result, the screening method further includes: determining the difference between the relevant read sequence and the reference sequence based on the alignment result through a mutation detection module to generate a mutation detection result.

Images