From: Strobealign: flexible seed size enables ultra-fast and accurate read alignment
Category | Tool | %-correct\(\uparrow\) | %-aligned\(\uparrow\) | Speed\(\uparrow\) | Mem\(\downarrow\) | Comment on overall results |
---|---|---|---|---|---|---|
Traditional | BWA-MEM | High | High | Low | Low | Works well in all benchmarked scenarios. High accuracy, but slow |
Bowtie2 | High | Medium | Low | Low | Works well in most benchmarked scenarios but is slow. Skips to align many reads for high diversity datasets | |
Recent | Strobealign | High | High | High | High | A very fast aligner with high accuracy. Peak performance for 150–500nt reads. Not recommended for 50–75nt reads |
BWA-MEM2 | High | High | Low to medium | High | Identical output to BWA-MEM at 1.2–2x speedup for paired-end alignment (our experiments), but very high memory usage | |
minimap2 | Medium | Medium | Medium | Medium | Relatively high accuracy on human genome, but not on repetitive genomes (rye and maize). Lower accuracy and slower than strobealign on most datasets but uses less memory | |
SNAP v2.0 | Medium | Low | Medium to high | High | Good accuracy and speed for low diversity datasets. Poor accuracy, speed, and percent aligned reads for high diversity and on the biological datasets | |
Accel-Align | Low | Medium | High | High | A very fast aligner but with lowest accuracy overall out of the benchmarked aligners |