Thread: subread
View Single Post
Old 06-04-2013, 11:04 PM   #7
Jon_Keats
Senior Member
 
Location: Phoenix, AZ

Join Date: Mar 2010
Posts: 279
Default

So by --mindist does this really mean the sequence length that is sequenced by both read1 and read2?

I only pick on this as it was a MAJOR issue in Tophat as read lengths increased and we started to cross-over resulting in negative values for their -inner-mate-distance. It might be worth creating a figure to illustrate the point. I'm happy to provide ours that outlines:
Library fragment
Insert Size
Inner-mate-distance

I would just add min-dist myself but I'm not confident I have it right or that we are completely talking about the same thing as I don't understand how you arrive at a paired-distance of 45 for a 146bp insert sequenced in both directions by 100bp, which results in bases 1-45 being sequenced by read1, bases 46-100 being sequenced twice (once by read1 and once by read2), and bases 101-146 being sequenced by read2. I would interpret -mindist as the distance separating the two reads, historically this was always a positive value as the reads never crossed each other. But now, particularly in RNAseq experiments and Agilent Exomes it is not at all uncommon that the two reads cross-over.
Jon_Keats is offline   Reply With Quote