ababaian / serratus

Ultra-deep search for novel viruses
http://serratus.io
GNU General Public License v3.0
254 stars 33 forks source link

Implement diamond for alignment step #151

Closed rcedgar closed 4 years ago

rcedgar commented 4 years ago

Diamond is much more sensitive to distant Covs than bowtie2, has similar execution speed and runs in well under 1Gb memory. See benchmark results in notebook/2000605_rce_diamond.

rcedgar commented 4 years ago

The container source is here including sumzer.py and Dockerfile:

s3://serratus-public/rce/mprot/container/

The reference mega-proteome and other miscellany are in this tarball, which the container fetches during load:

mprot.tz

The reference is out/mprot.fa in the tarball.

This is all very preliminary and -- please note -- subject to change.

rcedgar commented 4 years ago

Pilot run accomplished, closing issue.