Mouse liver circadian factor binding audit

Circadian ChIP-seq reproducibility audit

A method reconstruction, sensitivity analysis, and local ENCODE-equivalent comparison for public mouse liver circadian factor ChIP-seq data.

What can be reproduced from the deposited data?

Summary

No tested condition reproduced both the deposited peak counts and peak sets.

The analysis used the public data and the method details available from the reference publication and GEO records. The full original protocol was not available. The local ENCODE-equivalent workflows were run locally and were not official DNAnexus executions.

Peak counts and overlap

Analysis and condition	CLOCK peaks	Overlap	BMAL1 peaks	Overlap
Reference deposited peaks	759	Reference	1,578	Reference
Reference publication parameters, soft-masked	174	165	607	593
Sensitivity: duplicate auto	295	242	935	738
Sensitivity: duplicate all	282	214	1,206	790
Sensitivity: model 5,30	279	210	1,206	790
Sensitivity: genome 1.87e9	279	213	1,135	783
Sensitivity: p-value 1e-4	656	358	3,493	1,204
Sensitivity: p-value 1e-3	2,895	568	13,869	1,494
MACS3 compatibility, soft-masked	128	56	361	313
Reference publication parameters, hard-masked	223	167	963	544
ENCODE-equivalent, hard-masked, p-value 1e-4	163	111	824	470
ENCODE-equivalent, hard-masked, p-value 1e-3	633	229	2,741	720

Overlap means deposited reference peaks that overlap the reconstructed peak set.

Processing diagrams

The diagrams separate reported steps from unreported steps and distinguish the local implementation from the ENCODE recommendations.

Reference-method reconstruction diagram — Reference-method reconstruction.

Local ENCODE-equivalent workflow diagram — ENCODE-equivalent recommendations and local implementation.

Peak interval interpretation diagram — ChIP-seq peaks are enrichment intervals. They are not individual motif calls or gene assignments.

Genome-wide peak distribution

Click any plot to open it in a new window. Then click the image again to zoom in.

These plots show factor-specific ChIP-seq enrichment intervals across mm9 chromosomes. A peak can contain a sequence motif, but motif scanning is a separate analysis. A peak is not assigned to a gene unless an explicit annotation rule is applied.

Reference-method reconstruction

CLOCK reference-method peak distribution across mm9 chromosomes — CLOCK deposited and reference-method intervals. The canonical *Clock* locus is marked on the chromosome 5 row.

BMAL1 reference-method peak distribution across mm9 chromosomes — BMAL1 deposited and reference-method intervals. The canonical *Arntl* locus, commonly called BMAL1, is marked on the chromosome 7 row.

Local ENCODE-equivalent workflow

CLOCK ENCODE-equivalent peak distribution across mm9 chromosomes — CLOCK deposited and hard-masked ENCODE-equivalent intervals at p-values 1e-4 and 1e-3.

BMAL1 ENCODE-equivalent peak distribution across mm9 chromosomes — BMAL1 deposited and hard-masked ENCODE-equivalent intervals at p-values 1e-4 and 1e-3.

Workflow record

Reference-method sensitivity workflow: 16 peak-calling tests.
Reference-method hard-mask workflow: canonical UCSC chromFaMasked.tar.gz.
Hard-masked ENCODE-equivalent thresholds: 1e-4 and 1e-3.
Raw sequencing files are not stored here. GEO and SRA accessions are recorded in the manifests.