View Single Post
Old 04-26-2014, 10:11 AM   #33
Brian Bushnell
Super Moderator
Location: Walnut Creek, CA

Join Date: Jan 2014
Posts: 2,707


For analysis of vertebrate RNA-seq data with Cufflinks, I do recommend all of those parameters:
maxindel=200000 intronlen=10 ambig=all xstag=unstranded xmtag=t

Although "ambig=all" or "ambig=random" would both be OK. The "intronlen", "xstag" and "xmtag" are things I added specifically for Cufflinks compatibility (or the tuxedo package in general) so they are not needed if you do downstream analysis with other tools. The only ones that affect the actual mapping are "ambig" and "maxindel". xstag and xmtag just add optional tags, and intronlen changes the "D" symbol in cigar strings to "N" for deletions longer than the specified length. There's no reason why any of them should be necessary, but for whatever reason, Cufflinks needs them.

Note that if you your data is strand-specific, and you know which protocol it was, then it's better to set "xstag=firststrand" or "xstag=secondstrand" than unstranded.

Last edited by Brian Bushnell; 04-26-2014 at 10:13 AM.
Brian Bushnell is offline   Reply With Quote