CAGI5 saturation MPRA variant effect correlation, HepG2 (Tang et al. 2025 Table 1)

Overview

Pearson correlation between predicted and MPRA-measured single-nucleotide variant effects; average Pearson r over three CREs (LDLR, SORT1, F9).

Consult the linked sources for architecture or protocol details. Missing evidence is not evidence of a missing capability.

14 recorded evaluations, 14 metric rows. A comparison chart has not yet been validated for these results. The table retains the individual findings and their sources.

View coverage and remaining gaps across all benchmarks

Results

Results are available, but no reviewed comparison panel is linked in this release.

All evaluations

14 evaluations · 14 results. Different protocols are not a single leaderboard.

Filter evaluations

Tested model or methodProtocolDatasetEvaluation setupMetricEvidence origin

Applied filters: All linked evaluations

Exact evaluated configurations and original reported results
Tested configuration	Protocol and dataset	Finding	Evidence and details
Configuration: CNN-GPN, LentiMPRA-embedding (Tang et al. 2025)	Protocol: CAGI5 saturation MPRA variant effect correlation, HepG2 (Tang et al. 2025 Table 1) Dataset: CAGI5 saturation mutagenesis MPRA, four CREs (as used by Tang et al. 2025)	0.332 pearson-correlation unitless · higher Uncertainty: Not reported by the source Coverage: Not reported scored / Not reported eligible	Author-reported evaluation · Source checked Methods, coverage and source CNN-GPN (LentiMPRA-embedding) on CAGI5 HepG2 regulatory-variant-20261009-protocol-tang2025-cagi5-hepg2 Aggregation: Not reported Evaluating the representational power of pre-trained DNA language models for regulatory genomics · Table 1 row 'LentiMPRA-embedding' / 'CNN-GPN', column 'HepG2'
Configuration: CNN-NT, LentiMPRA-embedding (Tang et al. 2025)	Protocol: CAGI5 saturation MPRA variant effect correlation, HepG2 (Tang et al. 2025 Table 1) Dataset: CAGI5 saturation mutagenesis MPRA, four CREs (as used by Tang et al. 2025)	0.185 pearson-correlation unitless · higher Uncertainty: Not reported by the source Coverage: Not reported scored / Not reported eligible	Author-reported evaluation · Source checked Methods, coverage and source CNN-NT (LentiMPRA-embedding) on CAGI5 HepG2 regulatory-variant-20261009-protocol-tang2025-cagi5-hepg2 Aggregation: Not reported Evaluating the representational power of pre-trained DNA language models for regulatory genomics · Table 1 row 'LentiMPRA-embedding' / 'CNN-NT', column 'HepG2'
Configuration: CNN-SEI, LentiMPRA-embedding (Tang et al. 2025)	Protocol: CAGI5 saturation MPRA variant effect correlation, HepG2 (Tang et al. 2025 Table 1) Dataset: CAGI5 saturation mutagenesis MPRA, four CREs (as used by Tang et al. 2025)	0.579 pearson-correlation unitless · higher Uncertainty: Not reported by the source Coverage: Not reported scored / Not reported eligible	Author-reported evaluation · Source checked Methods, coverage and source CNN-SEI (LentiMPRA-embedding) on CAGI5 HepG2 regulatory-variant-20261009-protocol-tang2025-cagi5-hepg2 Aggregation: Not reported Evaluating the representational power of pre-trained DNA language models for regulatory genomics · Table 1 row 'LentiMPRA-embedding' / 'CNN-SEI', column 'HepG2'
Configuration: CNN, LentiMPRA-one-hot (Tang et al. 2025)	Protocol: CAGI5 saturation MPRA variant effect correlation, HepG2 (Tang et al. 2025 Table 1) Dataset: CAGI5 saturation mutagenesis MPRA, four CREs (as used by Tang et al. 2025)	0.324 pearson-correlation unitless · higher Uncertainty: Not reported by the source Coverage: Not reported scored / Not reported eligible	Author-reported evaluation · Source checked Methods, coverage and source CNN (LentiMPRA-one-hot) on CAGI5 HepG2 regulatory-variant-20261009-protocol-tang2025-cagi5-hepg2 Aggregation: Not reported Evaluating the representational power of pre-trained DNA language models for regulatory genomics · Table 1 row 'LentiMPRA-one-hot' / 'CNN', column 'HepG2'
Configuration: MPRAnn, LentiMPRA-one-hot (Tang et al. 2025)	Protocol: CAGI5 saturation MPRA variant effect correlation, HepG2 (Tang et al. 2025 Table 1) Dataset: CAGI5 saturation mutagenesis MPRA, four CREs (as used by Tang et al. 2025)	0.381 pearson-correlation unitless · higher Uncertainty: Not reported by the source Coverage: Not reported scored / Not reported eligible	Independent external evaluation · Source checked Methods, coverage and source MPRAnn (LentiMPRA-one-hot) on CAGI5 HepG2 regulatory-variant-20261009-protocol-tang2025-cagi5-hepg2 Aggregation: Not reported Evaluating the representational power of pre-trained DNA language models for regulatory genomics · Table 1 row 'LentiMPRA-one-hot' / 'MPRAnn', column 'HepG2'
Configuration: Residualbind, LentiMPRA-one-hot (Tang et al. 2025)	Protocol: CAGI5 saturation MPRA variant effect correlation, HepG2 (Tang et al. 2025 Table 1) Dataset: CAGI5 saturation mutagenesis MPRA, four CREs (as used by Tang et al. 2025)	0.485 pearson-correlation unitless · higher Uncertainty: Not reported by the source Coverage: Not reported scored / Not reported eligible	Author-reported evaluation · Source checked Methods, coverage and source Residualbind (LentiMPRA-one-hot) on CAGI5 HepG2 regulatory-variant-20261009-protocol-tang2025-cagi5-hepg2 Aggregation: Not reported Evaluating the representational power of pre-trained DNA language models for regulatory genomics · Table 1 row 'LentiMPRA-one-hot' / 'Residualbind', column 'HepG2'
Configuration: GPN (human), Self-supervised pre-training (Tang et al. 2025)	Protocol: CAGI5 saturation MPRA variant effect correlation, HepG2 (Tang et al. 2025 Table 1) Dataset: CAGI5 saturation mutagenesis MPRA, four CREs (as used by Tang et al. 2025)	0.002 pearson-correlation unitless · higher Uncertainty: Not reported by the source Coverage: Not reported scored / Not reported eligible	Independent external evaluation · Source checked Methods, coverage and source GPN (human) (Self-supervised pre-training) on CAGI5 HepG2 regulatory-variant-20261009-protocol-tang2025-cagi5-hepg2 Aggregation: Not reported Evaluating the representational power of pre-trained DNA language models for regulatory genomics · Table 1 row 'Self-supervised pre-training' / 'GPN (human)', column 'HepG2'
Configuration: HyenaDNA, Self-supervised pre-training (Tang et al. 2025)	Protocol: CAGI5 saturation MPRA variant effect correlation, HepG2 (Tang et al. 2025 Table 1) Dataset: CAGI5 saturation mutagenesis MPRA, four CREs (as used by Tang et al. 2025)	0.064 pearson-correlation unitless · higher Uncertainty: Not reported by the source Coverage: Not reported scored / Not reported eligible	Independent external evaluation · Source checked Methods, coverage and source HyenaDNA (Self-supervised pre-training) on CAGI5 HepG2 regulatory-variant-20261009-protocol-tang2025-cagi5-hepg2 Aggregation: Not reported Evaluating the representational power of pre-trained DNA language models for regulatory genomics · Table 1 row 'Self-supervised pre-training' / 'HyenaDNA', column 'HepG2'
Configuration: NT (2B51000G), Self-supervised pre-training (Tang et al. 2025)	Protocol: CAGI5 saturation MPRA variant effect correlation, HepG2 (Tang et al. 2025 Table 1) Dataset: CAGI5 saturation mutagenesis MPRA, four CREs (as used by Tang et al. 2025)	0.125 pearson-correlation unitless · higher Uncertainty: Not reported by the source Coverage: Not reported scored / Not reported eligible	Independent external evaluation · Source checked Methods, coverage and source NT (2B51000G) (Self-supervised pre-training) on CAGI5 HepG2 regulatory-variant-20261009-protocol-tang2025-cagi5-hepg2 Aggregation: Not reported Evaluating the representational power of pre-trained DNA language models for regulatory genomics · Table 1 row 'Self-supervised pre-training' / 'NT (2B51000G)', column 'HepG2'
Configuration: NT (2B5Species), Self-supervised pre-training (Tang et al. 2025)	Protocol: CAGI5 saturation MPRA variant effect correlation, HepG2 (Tang et al. 2025 Table 1) Dataset: CAGI5 saturation mutagenesis MPRA, four CREs (as used by Tang et al. 2025)	0.112 pearson-correlation unitless · higher Uncertainty: Not reported by the source Coverage: Not reported scored / Not reported eligible	Independent external evaluation · Source checked Methods, coverage and source NT (2B5Species) (Self-supervised pre-training) on CAGI5 HepG2 regulatory-variant-20261009-protocol-tang2025-cagi5-hepg2 Aggregation: Not reported Evaluating the representational power of pre-trained DNA language models for regulatory genomics · Table 1 row 'Self-supervised pre-training' / 'NT (2B5Species)', column 'HepG2'
Configuration: NT (500M1000G), Self-supervised pre-training (Tang et al. 2025)	Protocol: CAGI5 saturation MPRA variant effect correlation, HepG2 (Tang et al. 2025 Table 1) Dataset: CAGI5 saturation mutagenesis MPRA, four CREs (as used by Tang et al. 2025)	0.041 pearson-correlation unitless · higher Uncertainty: Not reported by the source Coverage: Not reported scored / Not reported eligible	Independent external evaluation · Source checked Methods, coverage and source NT (500M1000G) (Self-supervised pre-training) on CAGI5 HepG2 regulatory-variant-20261009-protocol-tang2025-cagi5-hepg2 Aggregation: Not reported Evaluating the representational power of pre-trained DNA language models for regulatory genomics · Table 1 row 'Self-supervised pre-training' / 'NT (500M1000G)', column 'HepG2'
Configuration: NT (500MHuman), Self-supervised pre-training (Tang et al. 2025)	Protocol: CAGI5 saturation MPRA variant effect correlation, HepG2 (Tang et al. 2025 Table 1) Dataset: CAGI5 saturation mutagenesis MPRA, four CREs (as used by Tang et al. 2025)	0.02 pearson-correlation unitless · higher Uncertainty: Not reported by the source Coverage: Not reported scored / Not reported eligible	Independent external evaluation · Source checked Methods, coverage and source NT (500MHuman) (Self-supervised pre-training) on CAGI5 HepG2 regulatory-variant-20261009-protocol-tang2025-cagi5-hepg2 Aggregation: Not reported Evaluating the representational power of pre-trained DNA language models for regulatory genomics · Table 1 row 'Self-supervised pre-training' / 'NT (500MHuman)', column 'HepG2'
Configuration: Enformer (DNase), Supervised one-hot (Tang et al. 2025)	Protocol: CAGI5 saturation MPRA variant effect correlation, HepG2 (Tang et al. 2025 Table 1) Dataset: CAGI5 saturation mutagenesis MPRA, four CREs (as used by Tang et al. 2025)	0.51 pearson-correlation unitless · higher Uncertainty: Not reported by the source Coverage: Not reported scored / Not reported eligible	Independent external evaluation · Source checked Methods, coverage and source Enformer (DNase) (Supervised one-hot) on CAGI5 HepG2 regulatory-variant-20261009-protocol-tang2025-cagi5-hepg2 Aggregation: Not reported Evaluating the representational power of pre-trained DNA language models for regulatory genomics · Table 1 row 'Supervised one-hot' / 'Enformer (DNase)', column 'HepG2'
Configuration: SEI, Supervised one-hot (Tang et al. 2025)	Protocol: CAGI5 saturation MPRA variant effect correlation, HepG2 (Tang et al. 2025 Table 1) Dataset: CAGI5 saturation mutagenesis MPRA, four CREs (as used by Tang et al. 2025)	0.545 pearson-correlation unitless · higher Uncertainty: Not reported by the source Coverage: Not reported scored / Not reported eligible	Independent external evaluation · Source checked Methods, coverage and source SEI (Supervised one-hot) on CAGI5 HepG2 regulatory-variant-20261009-protocol-tang2025-cagi5-hepg2 Aggregation: Not reported Evaluating the representational power of pre-trained DNA language models for regulatory genomics · Table 1 row 'Supervised one-hot' / 'SEI', column 'HepG2'

Source checking is not independent reproduction. Release 2026-10-10-7fcc3e48a123.

Methods and evaluation design

Procedure, tasks and evaluated configurations

Recorded evaluations

Each evaluation records what was tested and under which conditions.

Explore all linked results

Baseline coverage

Reference methods help show what a model adds beyond simple controls. We track a null control and a conventional method for each protocol.

0 of 2 active baseline roles have published Rewire measurements in this release. Measurements on a selected protocol do not establish coverage of an entire suite.

No execution recipe linked to this protocol. Recipe availability does not establish a completed evaluation.

Author-reported evaluations: 5
External evaluations: 9

Literature evidence is not a Rewire measurement. Executed but unpublished runs and private review status are not included.

Null control

Proposed control: requires review

Select a task-valid null control after reviewing inputs and metric

Protocol-specific applicability, permitted inputs, access, split, evaluator and execution requirements need review before implementation or execution.

This is a suggested selection rule, not a validated method or a measured score.

Conventional reference

Proposed control: requires review

Select an upstream conventional reference after reviewing the full protocol

Protocol-specific applicability, permitted inputs, access, split, evaluator and execution requirements need review before implementation or execution.

This is a suggested selection rule, not a validated method or a measured score.

Protocol coverage CSV (gzip) · Model evaluation matrix (gzip) · Source table (gzip) · Release and checksums (gzip)

Coverage is derived from release 2026-10-10-7fcc3e48a123. Source citations describe the original records; they do not validate an unreviewed baseline proposal. No results have been generated by this audit.

Run instructions

No runnable recipe has been reviewed for this protocol. Dataset access, model requirements, licences and compute requirements must be checked against its sources before execution.

Evidence

Source checking verifies the cited claim or transcription. It does not establish independent reproduction.

Evidence table

Inspect claims, sources and review details

Trace each statement to its source and review. A context-only reference supports the record generally; it does not verify an individual field. Source checking does not reproduce an experiment.

One row per statement and cited source. Multiple citations are not independent evaluations. Shared locators are labelled explicitly.

0 evidence rows matching the loaded filters

Claims, original sources and review scope · Release 2026-10-10-7fcc3e48a123
Property and statement	Original source and location	Review and provenance

No evidence rows match these filters. Choose another scope or clear the search.

All evidence (JSONL) (gzip)How to read the evidence

Sources and history

Release 2026-10-10-7fcc3e48a123 · Record review: source checked

1 source records and release history

Evaluating the representational power of pre-trained DNA language models for regulatory genomics · Original source · Genome Biology 26:203, published 2025-07-14; PMC12261763 full-text XML

Download this release (gzip)

Technical metadata and extraction receipts

Stable ID: regulatory-variant-20261009-protocol-tang2025-cagi5-hepg2

areas: dna-genomes
contexts: research
protocol: Zero-shot or transfer variant effect scores compared with experimental saturation mutagenesis effect sizes per CRE by Pearson correlation.
version: Table 1
source locator: Table 1 column 'HepG2' and footnote; Methods 'CAGI dataset'
limitations: Reporter (MPRA) activity on episomal constructs, not endogenous regulation.; Very small: one element (PKLR) in K562 and the average over three elements in HepG2, printed as 'LDLT' (presumably LDLR), SORT1 and F9; no uncertainty printed.; The lentiMPRA-trained probes, CNN and ResidualBind are the authors' own models (author_reported); MPRAnn is a published architecture retrained by the authors; the zero-shot language models, Sei and Enformer are published models.; Overlap between the CAGI5 elements and the lentiMPRA training sequences is not stated.

Related records

uses data: CAGI5 saturation mutagenesis MPRA, four CREs (as used by Tang et al. 2025)
subject: scope_note: regulatory-variant-20261009-protocol-tang2025-cagi5-hepg2
assessment: CNN-GPN (LentiMPRA-embedding) on CAGI5 HepG2
assessment: CNN-NT (LentiMPRA-embedding) on CAGI5 HepG2
assessment: CNN-SEI (LentiMPRA-embedding) on CAGI5 HepG2
assessment: CNN (LentiMPRA-one-hot) on CAGI5 HepG2
assessment: MPRAnn (LentiMPRA-one-hot) on CAGI5 HepG2
assessment: Residualbind (LentiMPRA-one-hot) on CAGI5 HepG2
assessment: GPN (human) (Self-supervised pre-training) on CAGI5 HepG2
assessment: HyenaDNA (Self-supervised pre-training) on CAGI5 HepG2
assessment: NT (2B51000G) (Self-supervised pre-training) on CAGI5 HepG2
assessment: NT (2B5Species) (Self-supervised pre-training) on CAGI5 HepG2
assessment: NT (500M1000G) (Self-supervised pre-training) on CAGI5 HepG2
assessment: NT (500MHuman) (Self-supervised pre-training) on CAGI5 HepG2
assessment: Enformer (DNase) (Supervised one-hot) on CAGI5 HepG2
assessment: SEI (Supervised one-hot) on CAGI5 HepG2
assessed by: Select regulatory variants and genes for functional follow-up