Population-scale STARR-seq of 3q25 DHS haplotypes in HepG2

The POP-STARR library captured haplotypes from 104 DNase-hypersensitive sites at the human 3q25 locus across 760 donors, using 174 custom amplicons and a population of reporter constructs. The pooled plasmid library was transfected into HepG2 cells in three replicate transfections; matched input-plasmid DNA and poly(A) RNA sequencing quantified haplotype regulatory activity.

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.

Organism: Human
Taxonomy ID: NCBITaxon:9606
Biosample: CVCL:0027
Reference genome: hg19
Design focus: Variant-focused
Region of interest: chr3:156781770-157036128

Perturbation & assay details

Basal / Untreated

POP-STARR is an episomal self-transcribing reporter assay in which population-derived regulatory haplotypes were cloned into the 3′ untranslated region of a STARR-seq reporter. The library was built from custom PCR-captured amplicons, pooled from eight donor pools, and transfected into HepG2 cells with Fugene HD. Three replicate transfections were harvested after approximately 48 hours; poly(A) RNA/cDNA and input plasmid DNA were sequenced, and haplotype activity was quantified from RNA-versus-DNA counts using Fisher’s exact tests with multiple-testing correction.

Quality control

The GEO release states that variants with read counts below 1 were discarded, haplotype activity was tested with Fisher’s exact tests comparing RNA and DNA counts, and P values were adjusted with R p.adjust() to control FDR. Package QC retained all 1,153 published haplotype-effect rows because every row had finite positive activity_effect, finite nonnegative absolute_log2_activity_effect, FDR in [0,1], a unique haplotype ID, a matching FASTA sequence, and a matching Table S2 amplicon coordinate. Non-significant measured haplotypes were retained; the 40 additional FASTA-only haplotypes without a published effect row were excluded.

Curation notes

The paper reports 104 DHSs, 174 amplicons, 1,153 unique haplotypes, and 942 variants. The public GEO FASTA contains 1,193 haplotype sequences across 172 amplicons, while the public effect table contains 1,153 rows across 171 amplicons; all effect rows matched a FASTA sequence and the processed table therefore excludes 40 FASTA-only records. The GEO family record has a legacy Series title mentioning preadipocytes, but its sample characteristics and the paper identify this assay as HepG2 POP-STARR. Transition/transversion counts were not recomputed from the VCF because the VCF contains additional calls, including indels, that are not represented in the haplotype-effect table; FASTA variant-state tokens are retained verbatim instead. Table S2 has no amplicon labels, so its row order was assigned to amp1–amp174, consistent with the numbered FASTA/effect records.

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

Population-scale STARR-seq of 3q25 DHS haplotypes in HepG2

Processed data

Quality control

Curation notes

Cite OpenMPRA

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

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

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