Skip to content

Performance

Andrea Telatin edited this page Feb 28, 2020 · 3 revisions

covtobed is a simple yet fast tool, several times faster than bedtools, and with comparable speed of mosdepth. In particular covtobed is faster when analysing small genomes, or large genomes having a very limited target (e. g. gene panels), while mosdepth is faster when analysing large genomes having larger targets (e. g. whole exome sequencing) or no target (e. g. whole genome sequencing).

When to use covtobed

covtobed is a simple utility that was built to be part of pipelines. You should consider covtobed if:

  • You want to reads from a standard input streaming
  • You want to process the produced BED with a pipe (standard output)
  • You want to use the pipeline both from Linux and macOS
  • You want to have a tool to do quick tests (e.g. bam files do not need to be indexed)

When analyzing large genomes it's conventient to compress and index the BED files. mosdepth provides a faster alternative and will directly compress and index the produced BED files, and carries in a single tools some feataures that can be useful in human genome resequencing analyses.

How to benchmark covtobed

We tested the speed using hyperfine, and we release the scripts used to compare the speed of bedtools, covtobed and mosdepth in different scenarios. As the results vary depending on the input file used and the operations needed, we recommend - if speed is an important factor - to perform a test on you actual datasets:

Clone this wiki locally