Dear DREAM Participants,   Alex Essebier (aesseb) noticed that some of the DNASE BAM files that we provided in Synapse did not appear to exactly match up with the provided DNASE peaks and bigwigs (See https://www.synapse.org/#!Synapse:syn6131484/discussion/threadId=605). We re-checked our uploads and found some inconsistencies that primarily affect some DNASE BAM files. Specifically, while the cell type identities of all files are correct, there are some replicate mismatches i.e. the BAM files correspond to one set of replicates whereas the peaks/bigwigs correspond to a different set. We have added, removed and modified some of the DNASE BAM files so that they exactly match the peak and bigwig sets. There was also one replicate mismatch in the DNASE bigwigs corresponding to the "induced_pluripotent_stem_cell" cell type. We've updated Synapse directories and tar archive files with the following changes. ### **Files REMOVED from "Challenge Data/DNASE/bams":** DNASE.IMR90.biorep1.techrep3.bam DNASE.IMR90.biorep1.techrep4.bam DNASE.HeLa-S3.biorep1.techrep1.bam DNASE.HepG2.biorep2.techrep2.bam DNASE.K562.biorep2.techrep11.bam ### **Files ADDED to "Challenge Data/DNASE/bams":** DNASE.A549.biorep1.techrep1.bam DNASE.A549.biorep1.techrep2.bam DNASE.A549.biorep1.techrep3.bam DNASE.GM12878.biorep1.techrep1.bam DNASE.H1-hESC.biorep1.techrep2.bam DNASE.HCT116.biorep1.techrep1.bam DNASE.Panc1.biorep1.techrep1.bam DNASE.SK-N-SH.biorep2.techrep1.bam ### **Files MODIFIED in "Challenge Data/DNASE/bams":** DNASE.H1-hESC.biorep1.techrep1.bam DNASE.IMR90.biorep1.techrep1.bam DNASE.IMR90.biorep1.techrep2.bam ### **Files MODIFIED in "Challenge Data/DNASE/fold_coverage_wiggles":** DNASE.induced_pluripotent_stem_cell.fc.signal.bigwig (Hence the "training_data.DNASE_wo_bams.tar" file in Challenge Data/ is also updated with this one update)   Sincere apologies for these errors and thank you for your patience. Also, thanks to Alex for identifying these inconsistencies.   **NOTE:** While these changes should not significantly affect your results since they are replicates, **PLEASE MAKE SURE YOU USE THESE LATEST FILES IN YOUR ANALYSES** so that everything is consistent. DNASE peak files are unaffected. Also none of the other data files are affected.   **NOTE:** As was previously stated in Data Description Section, all provided replicate BAM files for a cell type were pooled and subsampled to a maximum of 50 million unique mapping reads before calling peaks and generating bigwig signals. This was done so as to even out any differences across cell types due to sequencing depth differences. So even though the replicates corresponding to the BAMs and peak/bigwigs now match up, for some cell types if you pooled reads from the all the BAM replicates, they would exceed 50 million reads. You are free to use all the reads in the BAM files.   Best, Anshul and Nathan on behalf of the organizers