| In bioinformatics,sequence alignment of biological genes has been a hot issue in scie ntific research.With the development of biology and the deepening of the study of species,the number of genetic data is increasing day by day.At the same time,a new generation of sequencing technology,sequencing time and cost reduction,high flux sequencing the s equence data increased significantly,an exponential growth trend,and every day thereare s till new biological sequence data were discovered and recorded,and the speed of data gen erated is far greater than the data processing speed,so the method of treatment on large s cale DNA the sequencing data need to be more efficient.It is of great significance to deal with massive data in a timely and effective manner,to reveal its meaning,to clarify and understand its biological significance,and to obtain useful information for human beings.Se quence alignment is the basic data for gene research,to determine the sequence of subseq uent similarity analysis,and the expression of predicted single nucleotide polymorphism(S NP)loci,alternative splicing analysis,correlation analysis and prediction of disease drug development,provide important information on functional genome project genome biology a nd identification of gene function,so as to improve sequence alignment speed is essential for sequencing data.By researching the sequence alignment algorithm,understand the current alignment sof tware on BWT(Burrows-Wheeler Transform)indexing technology based on high accuracy and small memory consumption,compared with other methods based on the hash table per formance,but there are many memory access times,larger consumption time.This article on how to improve the alignment speed BWT indexing technology research,put forward a n application method of single thread parallel optimization algorithm improved from the ori ginal internal function operation way AVX(Advanced vector Extensions)instruction techn ology Intel micro architecture based on the function of large amount of calculation,many recursive calls to the BWT algorithm in rewriting,reduce the number of memory accesses to improve the algorithm execution speed,thus speeding up the processing speed of sequen ce alignment,BWT index technology time consumption and space consumption balance,i mprove the speed of the sequence alignment.In this paper,we design and implement a BWT sequence alignment algorithm based o n the AVX instruction set optimization.After studying the code and the implementation pr ocess of the BWT algorithm,we find the part that is suitable for AVX code rewriting.TheOCC function is calculated and results in memory,calculated using the AVX instruction se t in the calculation of the calculation,so that the computer can execute a command of mu ltiple parallel computing,greatly shorten the running time of the algorithm.In order to veri fy the accuracy and speed of the comparison results of the ratio,the algorithm of soybean sequencing data,the experimental results show that the proposed improved instruction can effectively improve the search efficiency of algorithm,accelerate the speed of sequence ali gnment.At the same time,the accuracy of the source code matching is improved by nearly50%,so that the time performance of the algorithm is significantly improved.Compared with the proposed BWT sequence alignment algorithm based on the AVX in struction set to the traditional algorithm,can effectively reduce the cycle of parallel compu ting search,traversal function calculation,calculate the number of CPU and reduce the nu mber of memory accesses and reduce the time complexity of the algorithm,improve the s peed of sequence alignment,the algorithm improves the performance of the time.And the performance of the algorithm is very stable,good performance in low disregistrated rate,h as practical significance for sequence alignment algorithm is more accurate,more efficientd ata analysis for gene sequence alignment method,to further accelerate the process of the whole genome sequence of the foundation. |
PDF Full Text Request