RT Journal Article
SR Electronic
T1 Accuracy, speed and error tolerance of short DNA sequence aligners
JF bioRxiv
FD Cold Spring Harbor Laboratory
SP 053686
DO 10.1101/053686
A1 Mark Ziemann
YR 2016
UL http://biorxiv.org/content/early/2016/05/16/053686.abstract
AB Aligning short DNA sequence reads to the genome is an early step in the processing of many types of genomics data, and impacts on the fidelity of downstream results. In this work, the accuracy, speed and tolerance to errors are evaluated in read of varied length for six commonly used mapping tools; BWA aln, BWA mem, Bowtie2, Soap2, Subread and STAR. The accuracy evaluation using Illumina-like simulated reads showed that accuracy varies by read length, but overall BWA aln was most accurate, followed by BWA mem and Bowtie2. BWA mem was most accurate with Ion Torrent-like read sets. STAR was at least 5 fold faster than Bowtie2 or BWA mem. BWA mem tolerated the highest density of mismatches and indels compared to other mappers. These data provide important accuracy and speed benchmarks for commonly used mapping software.