H3K27ac ChIP-seq QC Report

Overview

All five stages pass library complexity QC. PBC1, PBC2, and NRF are in the ideal tier for every ChIP and input library. Library bottlenecking is not a concern for any sample.

mHEP signal quality is reduced. RSC falls below 1 (0.966) and IDR-optimal peak count is the lowest of any stage (8,197), consistent with downstream signal and reproducibility issues rather than library complexity problems. NSC is also sub-threshold (<1.05) for ESC, HB, and mHEP — typical for broad histone-mark ChIP but worth noting.

Stages

ESC · DE · HB · iHEP · mHEP

Libraries

5 rep1 + 5 ctl1

Complexity flags

0 / 10

All in ideal tier

Reproducibility (IDR)

5 / 5

All pass (single-rep SCR)

Mapping rate

≥ 98.8%

Across all libraries

RSC flag (mHEP rep1)

0.966

Below ENCODE threshold of 1.0

Alignment

Bowtie2 alignment to hg38. Duplicate rates from Picard MarkDuplicates. Usable reads = post-dedup properly-paired fragments used for peak calling.

Stage	Library	Raw reads	Mapped %	Dup %	Usable reads
ESC	rep1	76,470,424	99.1%	2.60%	67,679,142
	ctl1	170,691,652	99.1%	1.98%	149,735,664
DE	rep1	80,810,794	98.8%	2.85%	70,563,206
	ctl1	175,384,390	98.9%	2.61%	151,262,582
HB	rep1	828,405,846	98.9%	3.52%	721,508,534
	ctl1	202,822,308	98.9%	2.34%	174,062,278
iHEP	rep1	114,992,366	98.9%	1.91%	102,455,270
	ctl1	206,416,690	99.0%	2.58%	176,428,360
mHEP	rep1	97,442,914	98.9%	2.18%	85,057,902
	ctl1	141,701,412	99.0%	2.23%	121,793,144

HB rep1 had an unusually large raw read count (828 M); usable reads after dedup remain well-matched to sequencing depth. All mapping rates and duplicate rates are within ENCODE-acceptable ranges.

Library Complexity

Flagging rule applied: orange if any of PBC1, PBC2, NRF falls in the orange tier; yellow if none orange but any yellow; none if all ideal. All ten libraries are comfortably in the ideal tier — no flags assigned.

Stage	Library	Total fragments	Distinct fragments	PBC1	PBC2	NRF	Flag
ESC	rep1	34,739,962	33,837,208	0.9740	38.34	0.9740	None
	ctl1	76,102,517	74,608,465	0.9803	50.75	0.9804	None
DE	rep1	36,314,719	35,279,470	0.9714	34.97	0.9715	None
	ctl1	77,455,203	75,437,964	0.9739	38.23	0.9740	None
HB	rep1	373,883,212	360,726,608	0.9647	28.26	0.9648	None
	ctl1	88,419,564	86,425,415	0.9774	44.18	0.9774	None
iHEP	rep1	52,221,795	51,224,110	0.9809	52.23	0.9809	None
	ctl1	89,438,106	87,328,510	0.9764	42.21	0.9764	None
mHEP	rep1	43,471,490	42,523,942	0.9782	45.81	0.9782	None
	ctl1	61,731,333	60,397,423	0.9784	46.11	0.9784	None

Interpretation: HB rep1 has the lowest PBC1 (0.965) and PBC2 (28.26) among ChIP libraries but is comfortably within the ideal tier (>0.9 / >10). mHEP rep1 (PBC1 0.978, PBC2 45.8) shows no bottlenecking — its downstream quality issues are not attributable to library complexity.

Signal Enrichment

Cross-correlation (NSC, RSC) computed on 15 M subsampled reads per replicate. JSD and FRiP computed on full usable-read BAMs. H3K27ac is a broad active-enhancer mark — NSC is expected to be lower than for sharp TF peaks.

Cross-correlation — NSC & RSC

Stage	Est. fragment (bp)	NSC	NSC flag	RSC	RSC flag
ESC	190	1.033	Orange	1.404	None
DE	200	1.057	Yellow	1.277	None
HB	225	1.048	Orange	1.281	None
iHEP	225	1.070	Yellow	1.379	None
mHEP	220	1.026	Orange	0.966	Yellow

JSD Enrichment & FRiP

Stage	JSD	% genome enriched	CH divergence	FRiP rep1 (SPP)	FRiP IDR-opt
ESC	0.142	39.2%	0.227	16.2%	1.6%
DE	0.190	33.6%	0.249	21.5%	4.3%
HB	0.269	24.6%	0.287	27.1%	15.7%
iHEP	0.328	23.7%	0.346	30.5%	9.2%
mHEP	0.168	34.1%	0.215	18.5%	1.1%

JSD: higher = greater divergence of ChIP vs. control (more enrichment). FRiP IDR-opt = fraction of rep1 reads in the IDR-optimal peak set — sensitive to IDR peak count.

mHEP enrichment context: RSC < 1 and an IDR-opt FRiP of 1.1% (vs 15.7% for HB and 9.2% for iHEP) are the clearest signal-quality flags for this stage. The low IDR peak count (8,197) drives the low IDR FRiP. NSC flags on ESC and HB are common for broad histone-mark ChIP and should be interpreted accordingly — RSC is the more robust indicator for these libraries.

Peak Calling & Reproducibility

Peak caller: SPP. Reproducibility by IDR and overlap on pseudo-replicates. Single biological replicate per stage → self-consistency ratio only (rescue ratio = 0, SCR = 1 by definition).

Stage	rep1 raw peaks	IDR-opt peaks	Overlap-opt peaks	IDR result	Overlap result
ESC	299,924	11,341	80,684	Pass	Pass
DE	299,933	23,990	122,865	Pass	Pass
HB	299,930	115,821	242,000	Pass	Pass
iHEP	299,922	48,862	192,065	Pass	Pass
mHEP	299,918	8,197	115,496	Pass	Pass

mHEP IDR peak count (8,197) is the lowest of all five stages and notably below the next-lowest (ESC: 11,341). The pipeline formally reports "pass" because in a single-replicate design the self-consistency ratio is 1.0 by construction — true biological reproducibility cannot be assessed without a second replicate. The low IDR count reflects poor pseudo-replicate concordance and is consistent with the sub-threshold RSC for this stage.

QC Thresholds Reference

ENCODE ChIP-seq pipeline v2 recommended thresholds used for badge assignment in this report.

PBC1 — PCR Bottleneck Coefficient 1

Orange< 0.5 (severe bottleneck)

Yellow0.5 – 0.8 (moderate bottleneck)

None0.8 – 0.9 acceptable; >0.9 ideal

PBC2 — PCR Bottleneck Coefficient 2

Orange< 1 (severe bottleneck)

Yellow1 – 3 (moderate bottleneck)

None3 – 10 acceptable; >10 ideal

NRF — Non-Redundant Fraction

Orange< 0.5 (concerning)

Yellow0.5 – 0.8 (moderate)

None0.8 – 0.9 acceptable; >0.9 ideal

NSC — Normalized Strand Correlation

Orange< 1.05

Yellow1.05 – 1.10

None> 1.10 (ideal)

Lower values are expected for broad histone marks (H3K27ac). Apply with context.

RSC — Relative Strand Correlation

Orange< 0.8

Yellow0.8 – 1.0

None> 1.0 (ideal)

FRiP — Fraction of Reads in Peaks

Orange< 1% (histone marks)

Yellow1% – 5%

None> 5% (ENCODE minimum for histone marks)