2-pass mapping - genome annotation of non model organism

54 views
Skip to first unread message

Andrei Rozanski

unread,
Nov 8, 2021, 2:13:13 PM11/8/21
to rna-star
Hello,

I am using RNASeq data to improve genome annotations of a non model organism. The so called hints for augustus can be extracted from the bam files and my question is about using 2-pass or not because the lack of annotations (--sjdbGTFfile). 
The manual says: "Using annotations is recommended either a the genome generation step, or mapping step."  
Is it advisable that I proceed with 1-pass only? I gave 2-pass a try for a few samples and it seems to improve the overall results. 

Sorry if I have missed this information elsewhere and thanks for the amazing job. 

All the best,
AndreiR

Alexander Dobin

unread,
Nov 12, 2021, 6:21:15 PM11/12/21
to rna-star
Hi Andrei,

the 2-pass is indeed helpful when you do not have annotations - it will allow to recover more spliced reads.
Effectively, the 1st pass creates an equivalent of annotated spliced junctions, and the 2nd pass uses them to improve mapping.

Cheers
Alex

Andrei Rozanski

unread,
Nov 15, 2021, 9:30:18 AM11/15/21
to rna-star
Dear Alex,

Thanks a lot for the reply. I will proceed with 2nd pass.

All the best,
AndreiR
Reply all
Reply to author
Forward
0 new messages