High-content single-cell combinatorial indexing

We describe a new generalizable library generation chemistry with increased efficiency that is amendable to tagmentation-based split-pool barcoding strategies, such as single-cell combinatorial indexing (sci). Symmetrical strand sci ('s3') uses a novel uracil-based adapter switching approach that provides an improved rate of conversion of source DNA into viable sequencing library fragments following tagmentation. We apply this chemistry to assay chromatin accessibility (s3-ATAC) to profile human cortical and mouse whole brain tissues, with mouse datasets demonstrating a 6-to-13-fold improvement in usable reads obtained per cell when compared to other available methods performed on the same sample type. We also demonstrate the generalizability of s3 by applying it to single-cell whole genome sequencing (s3-WGS), and whole genome plus chromatin conformation (s3-GCC), for structural variant calling in a patient-derived cancer cell line model. Using the high-coverage profiles produced by the s3 technologies we characterized preserved clonal structure and identified a putative subclone-specific translocation. Overall design: s3-ATAC was performed on two samples, a flash-frozen mouse whole brain and a flash-frozen human cortex sample. Library generation was performed as a barnyard with an equimolar mix of human and mouse nuclei at various steps. Cross-species nuclei were mixed either prior to tagmentation, after tagmentation, or kept isolated. See Fig2a in our manuscript or metadata files for more details. For s3-WGS and s3-GCC data, was performed on patient derived cell lines (PDCLs) and GM12878. Processed files are supplied. s3-WGS data is supplied as a read count per bin (s3-GCC read count data is included) as well as a bed format of read alignment location. s3-GCC data is supplied as a pairix format file with distal contacts information.