Experiment / E7E1EF83GEpisomal Plasmid MPRA

MPRA stage 2: replication, new variants, and oligo-configuration controls

Functional regulatory variants implicate distinct transcriptional networks in dementia

MPRA2 used the same 162-bp episomal HEK293T reporter library to re-test selected MPRA1 SigVars and dropouts, negative controls, orientation/position controls, and 483 additional variants from 11 Alzheimer's disease and four progressive supranuclear palsy loci. Six RNA libraries and four plasmid-DNA libraries are represented in the public mapped-count matrix.

Download processed table (CSV)Raw source data API documentation

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:0063
Reference genome: hg19
Design focus: Variant-focused
Region of interest: Not reported / not applicable

Perturbation & assay details

Basal / Untreated

MPRA2 used the same custom eGFP/minimal-promoter plasmid design, with 162-bp allele-containing oligos upstream of the promoter and a 20-bp barcode in the eGFP 3' UTR. The library included newly selected and missing variants plus replicated SigVars, negative controls, reverse-complement, lower-third/upstream, and reverse-orientation constructs. The released matrix contains six HEK293T RNA replicates (HEK1-HEK6) and four plasmid-DNA replicates (Plasmid1-Plasmid4); the paper describes six biological replicate transfections, while the DNA matrix has four DNA aliquots.

Processed data

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

Visible columns (38 of 38)

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

Page 1 · 50 rows · More results available

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

Column dictionary · 38 definitions

element_id: Stable package identifier for the MPRA2 variant-construct pair.
variant_id: Authors' internal numeric MPRA variant identifier; repeated across construct configurations when applicable.
rsid: dbSNP rs identifier reported by the authors.
chromosome: Chromosome reported in the author metadata.
position_hg19: Variant coordinate on hg19/GRCh37.
reference_allele: Reference allele (A0) represented by the reference oligo.
alternate_allele: Alternate allele (A1) represented by the alternate oligo.
regulomedb: RegulomeDB annotation supplied by the authors.
construct_type: Author-supplied MPRA2 construct category: new, missing, negative, fdr, RC, upstream, or reverse.
construct_description: Description of the MPRA2 screening, replication, control, orientation, or position configuration.
ref_oligo_name: Raw reference-allele element name in the GEO count matrix.
alt_oligo_name: Raw alternate-allele element name in the GEO count matrix.
ref_oligo_sequence: Actual 162-bp reference insert recovered from the GEO 210-mer library sequence list.
alt_oligo_sequence: Actual 162-bp alternate insert recovered from the GEO 210-mer library sequence list.
ref_qc_barcode_count: Number of reference barcodes passing the per-barcode count filter.
alt_qc_barcode_count: Number of alternate barcodes passing the per-barcode count filter.
ref_dna_total_raw_count: Sum of raw reference barcode counts across the four plasmid-DNA replicates after barcode QC.
alt_dna_total_raw_count: Sum of raw alternate barcode counts across the four plasmid-DNA replicates after barcode QC.
ref_rna_total_raw_count: Sum of raw reference barcode counts across the six HEK293T RNA replicates after barcode QC.
alt_rna_total_raw_count: Sum of raw alternate barcode counts across the six HEK293T RNA replicates after barcode QC.
ref_dna_mean_raw_count: Mean aggregate reference barcode count per plasmid-DNA replicate.
alt_dna_mean_raw_count: Mean aggregate alternate barcode count per plasmid-DNA replicate.
ref_rna_mean_raw_count: Mean aggregate reference barcode count per RNA replicate.
alt_rna_mean_raw_count: Mean aggregate alternate barcode count per RNA replicate.
ref_dna_mean_cpm: Mean reference plasmid-DNA CPM across DNA replicates after pseudocount normalization.
alt_dna_mean_cpm: Mean alternate plasmid-DNA CPM across DNA replicates after pseudocount normalization.
ref_rna_mean_cpm: Mean reference RNA CPM across HEK293T RNA replicates after pseudocount normalization.
alt_rna_mean_cpm: Mean alternate RNA CPM across HEK293T RNA replicates after pseudocount normalization.
ref_activity_median_log2_rna_dna: Median reference barcode log2 RNA/DNA activity after quantile normalization.
alt_activity_median_log2_rna_dna: Median alternate barcode log2 RNA/DNA activity after quantile normalization.
mpra_log2_effect: Alternate activity minus reference activity in log2 RNA/DNA units.
mpra_p_value: Two-sided tie-corrected normal-approximation Mann-Whitney U p-value for alternate versus reference barcode activity.
mpra_fdr: Benjamini-Hochberg FDR recalculated across QC-passing MPRA2 construct/variant rows in this package.
mpra_significant_fdr_0_01: Boolean indicating whether mpra_fdr is below 0.01.
mpra_effect_direction: Direction of the alternate-minus-reference activity effect.
rna_replicate_count: Number of HEK293T RNA replicates used (six).
dna_replicate_count: Number of plasmid-DNA replicates used (four).
source_table: Public source count matrix and author variant metadata used for the row.

Quality control

The authors' barcode-level QC was followed: retain a barcode only if it has at least 1 count in every RNA replicate and at least 5 counts in every plasmid-DNA replicate; add a pseudocount of 1, normalize each library to CPM, calculate log2 RNA/DNA ratios using the median plasmid-DNA CPM, and quantile-normalize the six RNA-ratio replicates. Variant-construct pairs with fewer than 5 retained barcodes for either allele were removed. The package retained 1,536 of 1,852 MPRA2 construct/variant pairs from 176,981 eligible barcode rows, corresponding to 3,072 retained allele elements. Allele activity is the median retained barcode ratio; the p-value is a two-sided Mann-Whitney U test using a tie-corrected normal approximation, and the reported FDR is a Benjamini-Hochberg recalculation across the packaged MPRA2 rows with significance marked at FDR < 0.01.

Curation notes

MPRA2 contains multiple construct configurations for some of the same underlying rsIDs. Rows are intentionally kept at the tested construct/variant level so RC, upstream, reverse, fdr, negative, new, and missing configurations remain distinguishable; they should not be treated as independent genomic variants in downstream analyses. The published methods describe four MPRA2 DNA replicates, consistent with the four Plasmid columns in GEO. The 1,536 passing construct/variant pairs correspond to 3,072 allele elements reported in the preprint's MPRA2 QC figure.

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