-
Notifications
You must be signed in to change notification settings - Fork 508
FAQ: Options and Parameters
- How to define maximum intron size?
- What is the best value of
--alignIntronMaxfor a particular species?
If --alignIntronMax is not specified, the maximum intron size (for most alignments) is defined by the alignment window" size
window = 2^(min(winBinNbits,genomeChrBinNbits))*winAnchorDistNbins
By default,
winBinNbits=16, genomeChrBinNbits=18 and winAnchorDistNbins=9
so alignment window = 589824
We need to specify --alignIntronMax to override this default alignment window size.
Relevant discussions:
https://github.com/alexdobin/STAR/issues/941#issuecomment-643680884
The actual value of --alignIntronMax for a particular species is determined by what your expected maximum intron size is. This expectation can come from the annotated junctions - for instance, in the human annotations, there are junctions with gaps ~1M bases, so I recommend --alignIntronMax 1000000. For smaller genomes, the maximum introns are usually smaller and so you would need to choose this value accordingly. Of course, you can always filter the gaps that are too long after mapping, so using a large value is fine.