After sequencing, FASTQ files were demultiplexed based upon the barcodes (cell type, replicate, DNA or cDNA) with “cutadapt” (Martin, 2011). Each sequencing read was mapped to an oligonucleotide sequence using “bwa-mem” (Li, 2013) and the number of unique molecular identifiers (UMI) for each sample was counted using “umi_tools” (Smith et al., 2017). We only counted reads with exact matches (edit distance (NM tag) equal to 0) to the oligo sequences of either reference or alternative alleles.
