Rough reduction algorithm for reduction of metagenomic DNA digital signature

The renewal of sequencing technology have promoted the rapid development of metagenomics, which have lead to the sharply increasing of sequenced metagenome DNA sequences number. But most of DNA sequences species are unknown and the current method is to extract nearly ten thousands digital features as the species label, and oversize feature dimension will make large amount of calculation, so how to reduce those large metagenome DNA digital features is a problem that bioinformatics faced. A rough reduction method based on Particle Swarm Optimization (PSO) algorithm is designed which take use of rough theory and combine PSO algorithm. This algorithm have been applied to reduce the metagenome DNA sequences on the level of genera which have approved that it can make effective reduction for metagenome DNA digital features and get more classification accuracy rate than before.