Per-read bead barcodes were parsed and trimmed using UMI-tools v1.0.041 (link). Constitutive elements of the bead barcodes were assigned to the closest known sequence allowing for up to 1 mismatch per 6-mer or 7-mer (mean >99% parsing efficiency across experiments). Paired-end reads were aligned to hg19 using BWA v0.7.1742 (link) on the Illumina BaseSpace online application. Bead-based ATAC-seq processing (BAP)17 (link) was used to identify systematic biases (i.e., reads aligning to an inordinately large number of barcodes) and barcode-aware deduplicate reads, as well as perform merging of multiple bead barcode instances associated with the same cell. Barcode merging was necessary due to the nature of the BioRad SureCell scATAC-seq procedure used in this study, which enables multiple beads per droplet. BAP was given an alignment (.bam) file for a given experiment with a bead barcode identifier indicated by a SAM tag as input. Aligned reads were combined using samtools merge (v1.9).
Free full text: Click here