Fig. 5.

Assessment of transcription factor footprint reproducibility across biological replicates for ATAC-seq and DNase1-seq. a ATAC-seq footprint biological reproducibility was investigated by overlapping the footprints of replicates and measuring the number of footprints with an overlapping replicate footprint. The percentage of overlap required between the two replicates was increased and the proportion of still-overlapping footprints plotted. For ATAC-seq data, the pipeline optimized to reproduce intermediate alignment and OCR files does not produce reproducible footprints as well as AUC-optimized pipelines. b Dnase1-seq footprint biological reproducibility was investigated by overlapping the footprints of replicates and measuring the number of footprints with an overlapping replicate footprint. The percentage of overlap required between the two replicates was increased and the proportion of still-overlapping footprints plotted. In DNase1-seq data, the pipeline optimized to reproduce intermediate alignment and OCR files produces more reproducible footprints than AUC-optimized or default pipelines. c The relationship between footprint reproducibility and ChIP-seq AUC was measured. Each pipeline’s proportion of footprints that overlap with the footprints of its replicate (at 1% overlap required) is plotted against that pipeline’s mean AUC. Pipelines with the highest AUC for each data type are marked with arrows