We quantified counts in peaks using the FeatureMatrix function in Signac and removed cells with <1,000 total counts. This function was parallelized via the future R package, allowing the user to determine the parallelization strategy used (https://cran.r-project.org/package=future). Signac also includes a convenience function (GenomeBinMatrix) to quantify signal in genomic bins tiling the entire genome or given chromosomes. Finally, we reduced the dimensionality by applying LSI and UMAP as described above, using LSI components 2 to 30.
