K562 large-scale lentiMPRA benchmark dataset
Designing DNA With Tunable Regulatory Activity Using Discrete DiffusionThe paper benchmarks D3 on a public K562 large-scale lentiMPRA library from Agarwal et al. (2025), consisting of genomic candidate enhancers, promoters, and controls measured in three biological replicates. This package contains the corresponding ENCODE processed element quantifications reused by the paper; it does not represent the paper’s separate 2,760-element synthetic-validation library.
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
Lentivirus-based genomic integration of 230-bp candidate regulatory elements upstream of a minimal promoter with the barcode in the 5' UTR; DNA and RNA barcode counts were processed with MPRAflow for three biological replicates. The raw table contains ENCODE-normalized element-level DNA and RNA counts, RNA/DNA ratios, log2 activity scores, and observed barcode counts.
Processed data
50 rows per page. Click a cell to inspect its full value.
Visible columns (27 of 27)
| 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 · 27 definitions
- element_id
- Exact ENCODE element identifier/name; reverse names preserve the source suffix.
- element_class
- Category inferred from the element name: candidate enhancer, promoter, dinucleotide-shuffled control, or other/control.
- chromosome
- GRCh38 chromosome from the associated BED file; blank when no coordinate was available.
- start_0based
- GRCh38 BED zero-based start coordinate.
- end_0based
- GRCh38 BED exclusive end coordinate.
- strand
- GRCh38 BED strand (+ or -).
- orientation
- Orientation derived from the BED strand: forward or reverse.
- replicate_count_passing
- Number of biological replicates passing the package QC filters.
- min_barcode_count
- Minimum observed barcode count across the three passing replicates.
- dna_count_rep1
- ENCODE normalized element DNA count for biological replicate 1.
- rna_count_rep1
- ENCODE normalized element RNA count for biological replicate 1.
- ratio_rep1
- ENCODE normalized RNA/DNA ratio for biological replicate 1.
- log2_activity_rep1
- ENCODE log2 RNA/DNA activity score for biological replicate 1.
- n_barcodes_rep1
- Observed barcode count (n_obs_bc) for biological replicate 1.
- dna_count_rep2
- ENCODE normalized element DNA count for biological replicate 2.
- rna_count_rep2
- ENCODE normalized element RNA count for biological replicate 2.
- ratio_rep2
- ENCODE normalized RNA/DNA ratio for biological replicate 2.
- log2_activity_rep2
- ENCODE log2 RNA/DNA activity score for biological replicate 2.
- n_barcodes_rep2
- Observed barcode count (n_obs_bc) for biological replicate 2.
- dna_count_rep3
- ENCODE normalized element DNA count for biological replicate 3.
- rna_count_rep3
- ENCODE normalized element RNA count for biological replicate 3.
- ratio_rep3
- ENCODE normalized RNA/DNA ratio for biological replicate 3.
- log2_activity_rep3
- ENCODE log2 RNA/DNA activity score for biological replicate 3.
- n_barcodes_rep3
- Observed barcode count (n_obs_bc) for biological replicate 3.
- log2_activity_mean
- Arithmetic mean of the three replicate log2 activity scores.
- log2_activity_sd
- Sample standard deviation of the three replicate log2 activity scores.
- qc_pass
- TRUE for rows retained after the package QC filters.
Quality control
The ENCODE MPRAflow combined table was filtered to remove the no_BC record; require all three biological replicates; require at least 10 observed barcodes (n_obs_bc) per replicate, positive normalized DNA and RNA counts, and finite ratio/log2 values. This retained 224,555 of 230,932 non-no_BC element IDs (all 3 replicates), and all retained rows have qc_pass=TRUE. GRCh38 BED coordinates were joined where available; missing coordinates did not cause exclusion.
Curation notes
The open v3 article reports a separate wet-lab K562 validation library of 2,760 elements and 2,534 retained elements, but its MPRAflow validation table is not included in the open supplement or linked Zenodo record. This package therefore represents the public ENCODE K562 large-scale lentiMPRA benchmark dataset explicitly reused by the paper (Agarwal et al. 2025), not the study’s new synthetic-vs-genomic validation results. The table retains exact ENCODE element names; reverse-orientation names include the _Reversed: suffix. dna_count/rna_count are the normalized count fields supplied by ENCODE, not raw FASTQ read counts. Aggregate activity SD is the sample standard deviation across the three replicates.