In Salmon transcripts disappear. Sailfish doesn't allow bias correction in Galaxy.

28 views
Skip to first unread message

Tatiana Krutikhina

unread,
May 7, 2020, 10:00:56 AM5/7/20
to Sailfish Users Group
Hello everyone,

I have an issue with aligners. I have a reference transcriptome (a long-read sequencing derived one), and illumina short-read sequencing data. I need to map the reads to reference transcriptome. I tried to do it with Salmon, first in Galaxy, and then in command-line version, and found out that some transcripts just disappered. They were just not present in the output.
I also tried Sailfish in Galaxy (Galaxy Version 0.10.1.1). It worked and transcripts didn't disappear, but if I added the length bias correction and GC bias correction, it output zero count file. Does it mean that the program works wrong? Can I trust Galaxy versions in general?

Thank you in advance.
Best regards,
Tanya

Rob

unread,
May 7, 2020, 10:53:04 AM5/7/20
to Sailfish Users Group
Hi Tanya,

  In general, we recommend using salmon instead of sailfish.

Regarding the problem of the transcripts disappearing; how many such transcripts vanished?  By default, salmon will collapse duplicate transcripts (transcripts with identical sequence) when indexing.  This results in only one from among the duplicate group being quantified.  The full information of what was collapsed (and what it was collapsed into) is retained in a file called `duplicate_clusters.tsv` in the index directory created by salmon.  This is a TSV file where the first column is the retained transcript (in the quant.sf file) and the second column is the transcript (identical) that was collapsed into it.  One retained transcript can have multiple duplicates.

Best,
Rob

Anand

unread,
May 7, 2020, 12:02:50 PM5/7/20
to Tatiana Krutikhina, Sailfish Users Group
I think this forum is no longer maintained




--
Sailfish is available at https://github.com/kingsfordgroup/sailfish
Citation:
Patro, Rob, Stephen M. Mount, and Carl Kingsford. "Sailfish enables alignment-free isoform quantification from RNA-seq reads using lightweight algorithms." Nature biotechnology 32.5 (2014): 462-464.
---
You received this message because you are subscribed to the Google Groups "Sailfish Users Group" group.
To unsubscribe from this group and stop receiving emails from it, send an email to sailfish-user...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/sailfish-users/3099cd7d-f674-4b47-9e1c-b88df08d874d%40googlegroups.com.

Anand

unread,
May 7, 2020, 12:04:29 PM5/7/20
to Tatiana Krutikhina, Sailfish Users Group
i guess I am wrong 

Rob responded here 

I always use the GH issue tracker.
Reply all
Reply to author
Forward
0 new messages