multimapping with STARlong

[Osnat Penn]

Hi,

I'm trying to use STARlong to get multimapping alignments using the outFilterMultimapScoreRange parameter.

However, it seems to be more strict than the regular STAR.

For example, I've tried to run both STAR and STARlong to align a 648bp transcript, using the same values for --outFilterMultimapScoreRange, --outFilterMismatchNmax, and --outFilterMultimapNmax.

STAR output includes dozens of secondary alignments, while with STARlong I only got the primary alignment (which shows 100% identity) and didn't get any secondary alignment.

Most of the transcripts that I wish to align are very long.

Is there a way to run STARlong to output all multimapping alignments ?

Thank you,

Osnat

[Alexander Dobin]

Hi Osnat,

STARlong normally examines fewer candidate loci, and thus finds fewer multimappers.

You can mitigate this by reducing the new (since 2.5.2a) parameter --winReadCoverageRelativeMin from 0.5 down to 0.3 or even 0.1 .

Please let me know how this goes as I have not tested this parameter extensively.

Cheers

Alex