Configurable Blast results filtering

ohnosequences / mg7

Configurable and scalable 16S metagenomics data analysis

https://goo.gl/y3rZFD

GNU Affero General Public License v3.0

3 stars 3 forks source link

Configurable Blast results filtering #38

Closed laughedelic closed 8 years ago

laughedelic commented 8 years ago

This is needed for Pacbio case.

Probably it should be added to the Blast step.

eparejatobes commented 8 years ago

The idea is to add to the configuration parameters a predicate on the BLAST output record values.

laughedelic commented 8 years ago

@eparejatobes @marina-manrique meeting notes:

For pacbio:

output fields should contain qcovs (query coverage for subject)
keeping only those for whom this is 100%
high evalue
filtering by a configurable predicate should be added to the blast step

In general:

filtering of result by qcovs is useful in general
max_target_seqs should be set to maximum/default
set as blast default 99% identity (perc_identity)

eparejatobes commented 8 years ago

qcovs should be always an output field; same about pident.
max_target_seqs should be as big as possible

laughedelic commented 8 years ago

Actually, I think this is done in #49. I made the filtering predicate configurable, and added the mentioned defaults in #51. Is there anything missing here?

eparejatobes commented 8 years ago

Fine for me

laughedelic commented 8 years ago

Leaving this open, due to the changes in #52: the predicate is not configurable for now (it caused problems).

laughedelic commented 8 years ago

In the end it's fixed in #52. Closing this.