Skip to first unread message
Ernest Szeto
May 9, 2014, 7:16:59 PM5/9/14
to img-use...@lbl.gov
Previously, the filters for assigning COG's via rpsblast was just the e-value <= 1e-2.
An additional criteria has been added to reduce false positives.
The length of the aligned region needs to be at least 70% of the COG consensus length.
As a result of this, expect the percentage of genes assigned to COG's to go down slightly,
but this is the reason for this.
An additional criteria has been added to reduce false positives.
The length of the aligned region needs to be at least 70% of the COG consensus length.
As a result of this, expect the percentage of genes assigned to COG's to go down slightly,
but this is the reason for this.
Ernest Szeto
Aug 7, 2014, 6:30:48 PM8/7/14
to img-use...@lbl.gov
In a few weeks, the same criteria of alignment percentage on the consensus sequence will also be applied to KOG's. This is due to reports of false positives or inconsistency with other annotations.
Reply all
Reply to author
Forward
0 new messages