Hi, I want to use vartrix on 5' 10X scRNA-seq data to genotype cells.
I have previously used cellsnp-lite but would like to compare with a more UMI aware tool that can get a consensus across reads of the same transcript.
When I compare coverage of cellsnp-lite results and vartrix results, I see differences at exon boundaries.
I assume it is due to the re-alignment of reads.
Below shown the UMI coverage of the CDS of a gene of interest. y-axis is --ref-matrix + --out-matrix which should represent transcript coverage at each position (not considering multi-allelic sites).
I have marked below where the depth of vartrix and cellsnp-lite differ (blue) and the exon boundaries. There are "dips" around the exon boundaries.
I'm concerned that I would miss many variants that affect splicing if re-alignment around exon boundaries is a problem.
Could you describe how you handle the paired-end data that CellRanger produces for 5' 10X data?
The read containing cell barcode and UMI also contains part of the transcript, so for each CB UMI pair, there are at least 2 reads covering distinct regions of the gene, with the same read name.
Below an example of a read pair from the CellRanger possorted_genome_bam.bam.
Hi, I want to use vartrix on 5' 10X scRNA-seq data to genotype cells.
I have previously used cellsnp-lite but would like to compare with a more UMI aware tool that can get a consensus across reads of the same transcript.
When I compare coverage of cellsnp-lite results and vartrix results, I see differences at exon boundaries.
I assume it is due to the re-alignment of reads.
Below shown the UMI coverage of the CDS of a gene of interest. y-axis is
--ref-matrix+--out-matrixwhich should represent transcript coverage at each position (not considering multi-allelic sites).I have marked below where the depth of vartrix and cellsnp-lite differ (blue) and the exon boundaries. There are "dips" around the exon boundaries.
I'm concerned that I would miss many variants that affect splicing if re-alignment around exon boundaries is a problem.
Could you describe how you handle the paired-end data that CellRanger produces for 5' 10X data?
The read containing cell barcode and UMI also contains part of the transcript, so for each CB UMI pair, there are at least 2 reads covering distinct regions of the gene, with the same read name.
Below an example of a read pair from the CellRanger
possorted_genome_bam.bam.Thanks a lot!
Best, Henrietta