Experiment / E08HMDWSQATAC-STARR-seq

GM12878 ATAC-STARR-seq accessible chromatin regulatory activity

ATAC-STARR-seq reveals transcription factor–bound activators and silencers within chromatin-accessible regions of the human genome

Three biological GM12878 replicate pairs (reisolated plasmid DNA and reporter RNA) assayed a Tn5-selected library of approximately 50 million unique accessible DNA fragments tiling approximately 101,000 hg38 accessible chromatin peaks. The processed table contains the study's nonredundant sliding-window active and silent regions, with the mean log2 RNA/DNA score and annotations to the deposited accessible peak set.

Processed tables are specific to each experiment. Column names, units, measurements, and table structure are not standardized across the database. Check this experiment’s column definitions and quality-control notes before comparing or combining data.

Perturbation & assay details

Basal / Untreated

Episomal STARR-seq plasmids use the bacterial origin of replication (Ori) as the minimal promoter; unique Tn5-selected accessible DNA inserts are cloned in the reporter 3′ UTR downstream of truncated GFP and a polyadenylation signal. Reporter RNA and reisolated plasmid DNA were harvested 24 hours after electroporation, and activity was quantified as the normalized log2 RNA/DNA fold-change. The same accessible-fragment library also supports chromatin-accessibility peak calling and Tn5 footprinting.

Processed data

50 rows per page. Click a cell to inspect its full value.

Visible columns (21 of 21)
Row
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50

Filters apply to this table only. The CSV download contains the complete processed table; filtered rows are available through the API.

Column dictionary · 21 definitions
element_id
Unique class-prefixed identifier for a processed regulatory region.
region_name
Original region name from the deposited active or silent BED file.
chrom
Chromosome in the hg38 reference assembly.
start
0-based inclusive genomic start coordinate.
end
0-based half-open genomic end coordinate.
length_bp
Region length in base pairs (end minus start).
strand
Strand field from the deposited BED call; '.' indicates unstranded.
activity_class
Regulatory activity class: activator or silencer.
log2_rna_dna_mean
Mean log2 fold-change of reporter RNA relative to reisolated plasmid DNA across merged significant 50-bp bins; positive values indicate activation and negative values indicate silencing.
accessible_peak_id
Genrich accessible-peak identifier with the greatest base-pair overlap with the region.
accessible_peak_count
Number of deposited Genrich accessible peaks overlapping the region.
best_peak_overlap_bp
Number of overlapping base pairs between the region and accessible_peak_id.
peak_start
0-based inclusive start coordinate of accessible_peak_id.
peak_end
0-based half-open end coordinate of accessible_peak_id.
peak_signal_value
Genrich narrowPeak signalValue for accessible_peak_id.
peak_pvalue_neglog10
Genrich narrowPeak -log10 p-value field for accessible_peak_id.
peak_qvalue_neglog10
Genrich narrowPeak -log10 q-value field for accessible_peak_id.
peak_summit_offset
Genrich narrowPeak summit offset relative to peak_start for accessible_peak_id.
opposite_activity_class
The opposite activity class when the region overlaps at least one deposited call of that class; blank otherwise.
opposite_activity_region_count
Count of deposited opposite-class regions overlapping this region.
source_file
Deposited GEO BED file from which the processed region call was derived.

Quality control

The paper's pipeline trimmed reads with Trim Galore, mapped them to hg38 with Bowtie2, removed reads with MAPQ <30 and reads mapping to mitochondrial DNA or ENCODE blacklist regions, and used DESeq2 on the three biological replicate pairs after removing bins with zero counts. Benjamini–Hochberg adjusted P <0.1 plus positive or negative log2 RNA/DNA fold-change defined active or silent bins; overlapping or book-ended bins were merged and their scores averaged. Activity analysis retained duplicates as specified by the authors; replicate correlations were high (Pearson r² 0.96–0.99; Spearman ρ 0.77–0.93). The processed table includes only the deposited 30,078 active and 21,125 silent merged calls and independently excludes malformed intervals or class-inconsistent scores (none were found).

Curation notes

This is a genome-wide accessible-chromatin screen rather than an allele-contrast or variant-focused MPRA. The GEO BED files have no headers; their score column was interpreted from the manuscript and author repository as the mean log2 fold-change (RNA/DNA) across merged significant sliding-window bins. The processed table selects the overlapping Genrich peak with greatest base-pair overlap (ties resolved by peak signal) and retains 50 records that overlap an opposite-class call. Raw FASTQ/SRA reads are not included per request; the GEO, SRA, and BioProject links are retained for retrieval. GM12878 is the EBV-transformed human B-cell line indexed by Cellosaurus CVCL:7526.

Cite OpenMPRA

Cite the OpenMPRA database. Include your access date because the collection changes over time.

Please also cite the source studies when using their data.