STARsolo: how to allow deduplicating UMIs based on start-end

[vivekbhr]

Hi @alexdobin . I want to count UMIs per gene per cell for Gene and Velocyto analysis, but based on start+end+umi per barcode, instead of geneID+umi per barcode which is what STAR seems to do. I tried de-duplicating the BAM myself, extract the fastq and map using, --soloUMIdedup NoDedup but it seems there's still de-duplication going on at least for Velocyto (spliced/unspliced) counts.

Do we have a way to deduplicate and count UMIs per gene based on position in STARsolo?

[alexdobin]

Hi Vivek,

the only possibility to collapse reads based on read start/end position is for SmartSeq data.
Presently there is no option to collapse based on both position and UMI.

[vivekbhr]

Hi @alexdobin thanks for the reply? Would it be easy to implement this option at some point? It would help with the kind of dataset I am working on

[vivekbhr]

Just saw the issue #1560 . If something like that is implemented, that would also work for me, as I can do the deduplication+counting myself if the reads are tagged as exon/intron/amb.