I used STAR 2.7.10a in a targeted scRNAseq experiment and noticed an issue with the gene CD99 from the targeted primer panel. Although STAR mapped CD99, it was not counted. Upon investigation, I found that CD99 maps to both chrX and chrY. After enabling the --soloMultiMappers Uniform feature, I observed that half of the CD99 UMIs were counted on chrX (CD99) and the other half on chrY (CD99.1).
Given that these multi-mapped reads essentially represent the same gene, it would make sense to collapse them and count them as one gene. I suspect there are many such multi-gene instances on chrX and chrY.
It would be beneficial to implement a feature to handle this specific case, separate from --soloMultiMappers. Because I am still not sure if I want to enable multimapping for multimapping cases that you mention in the manual: reads that map uniquely to a genomic region where two or more genes overlap, or reads that map to multiple loci in the genome, with each locus annotated to a different gene.
Hi,
I used STAR 2.7.10a in a targeted scRNAseq experiment and noticed an issue with the gene CD99 from the targeted primer panel. Although STAR mapped CD99, it was not counted. Upon investigation, I found that CD99 maps to both chrX and chrY. After enabling the --soloMultiMappers Uniform feature, I observed that half of the CD99 UMIs were counted on chrX (CD99) and the other half on chrY (CD99.1).
Given that these multi-mapped reads essentially represent the same gene, it would make sense to collapse them and count them as one gene. I suspect there are many such multi-gene instances on chrX and chrY.
It would be beneficial to implement a feature to handle this specific case, separate from --soloMultiMappers. Because I am still not sure if I want to enable multimapping for multimapping cases that you mention in the manual: reads that map uniquely to a genomic region where two or more genes overlap, or reads that map to multiple loci in the genome, with each locus annotated to a different gene.
What do you think?
Thanks!