An Exact Matching Method for 16S rRNA Taxonomy Classification

One popular approach to taxonomy classification in the microbiome utilizes 16S ribosomal RNA sequences. The main challenge is that 16S rRNA sequences could be almost identical in closely related species, and it is difficult to distinguish them at the species level. Recent approaches are able to achieve almost single nucleotide resolution by constructing an error model of the reads. We develop an exact matching algorithm to utilize the single nucleotide resolution directly. We show that our algorithm is able to obtain improved accuracy in recent samples of mock communities and in samples of high compositional complexity when compared to existing algorithms. A software program implementing this algorithm is available at http://faculty.cse.tamu.edu/shsze/kmpmatch.