To use the 10x dataset through my SNV calling pipeline, I used the BAM file from the 10x datasets downloading page. The only issue is that the 10x BAM file pull all the file together. So I used a script to "demultiplex" the bam file and create fastq files for each cell in different folders. Then, in a second time, I re-align with STAR and used my pipeline to call SNVs. I have added the script to the git rep here: https://github.com/lanagarmire/SSrGE/blob/master/garmire_SNV_calling/parse_10x_bam_file_to_fastq_files.py
Calling SNVs from scRNA-Seq 10x datasets was relatively successful for us because we were
able to obtain a true positive rate of 0.65 (+/- 5) using 10x reads when injecting 50K simulated SNVs in the exonic regions if at least 3 reads cover the SNV position. These SNVs were sufficient to separate subpopulations.
To use the 10x dataset through my SNV calling pipeline, I used the BAM file from the 10x datasets downloading page. The only issue is that the 10x BAM file pull all the file together. So I used a script to "demultiplex" the bam file and create fastq files for each cell in different folders. Then, in a second time, I re-align with STAR and used my pipeline to call SNVs. I have added the script to the git rep here: https://github.com/lanagarmire/SSrGE/blob/master/garmire_SNV_calling/parse_10x_bam_file_to_fastq_files.py
Calling SNVs from scRNA-Seq 10x datasets was relatively successful for us because we were able to obtain a true positive rate of 0.65 (+/- 5) using 10x reads when injecting 50K simulated SNVs in the exonic regions if at least 3 reads cover the SNV position. These SNVs were sufficient to separate subpopulations.