Applying to 10x scRNAseq data

[amitlakhanpal]

Hello! Thank you for creating this tool.

I came to it from one of the 10x packages scHLAcount (https://github.com/10XGenomics/scHLAcount) that recommends arcasHLA as a superior option for HLA genotyping.

Would like to check with the experts here if the approach I've used (derived from the comment in #49 to invoke the routine as for single-read) is appropriate, and suggest that it might be added to the documentation if it is, since this might be a somewhat common usecase for people using the 10x system.

./arcasHLA extract --single --unmapped --o $outputdirectory -t 8 -v /path/to/10x/possorted_genome_bam.bam

./arcasHLA genotype --single -g A,B,C,DPB1,DQB1,DRB1 -o $outputdirectory -t 8 -v $outputdirectory/possorted_genome_bam.extracted.fq.gz

Also, with apologies if I've missed this in documentation somewhere, a few related quetions:
(1) is there a usual metric of confidence in the typing assignments that is generated by arcas?
(2) is it safe to say that the reason for using the single-end flag despite the protocol doing paired-end reads is that the 10x method uses one of the reads essentially just for the barcode and UMI (as in https://www.seqanswers.com/forum/sequencing-technologies-companies/illumina-solexa/65300-10x-genomics-and-paired-end-reads)?
(3) would it be likely that there might be a more optimized/streamlined application of arcas to 10x scRNAseq data?

Thanks again!