Fast gapped-read alignment with Bowtie 2

Abstract

As the rate of sequencing increases, greater throughput is demanded from read aligners. The full-text minute index is often used to make alignment very fast and memory-efficient, but the approach is ill-suited to finding longer, gapped alignments. Bowtie 2 combines the strengths of the full-text minute index with the flexibility and speed of hardware-accelerated dynamic programming algorithms to achieve a combination of high speed, sensitivity and accuracy.

Figure 1

Alignment comparison using HiSeq 2000, 454 and Ion Torrent reads. (a–d) Bowtie 2, BWA, SOAP2 and Bowtie were used to align two million 100 nt × 100 nt paired-end HiSeq 2000 reads from a resequencing study. Shown are results for unpaired alignment of end 1 (a), paired-end alignment (b), Bowtie 2 and BWA-SW alignment of 1 million 454 reads from the 1000 Genomes Project Pilot (c), and Bowtie 2 and BWA-SW to align one million Ion Torrent reads from the G. Moore resequencing project (d). Plotted is the percentage of reads for which at least one alignment was found. Each numbered point is data obtained using command-line parameters shown in Supplementary Table 1.

Figure 2

Sensitivity and accuracy of alignment using simulated reads. (a–c) Cumulative number of correct and incorrect alignments from high to low mapping quality for simulated Illumina-like unpaired 100 nt and 150 nt datasets (a), for simulated Illumina-like paired-end 100 nt × 100 nt and 150 nt × 150 nt datasets (b), and for simulated 454-like datasets with average read lengths 250 nt and 400 nt (c) using indicated aligners.