Strengths and considerations
No source-reviewed explanatory claims are recorded here yet.
benchmark · protocol
Supervised multimodal CAGI5 MPRA prediction (AlphaGenome paper)
Does supervised combination of sequence-derived modalities better predict CAGI5 reporter-assay effects?
AlphaGenome Nature 2026 supplementary comparison tables; AlphaGenome Nature 2026 supplementary methods; alphagenome: Journal full-text XML · 'Suppl Table 4 Variant performan'!A12:P12; Supplementary Methods p.30, Chromosome Splits for Variant Benchmarks; Supplementary Methods p.40, Benchmarking on CAGI5 MPRA; Table10 cell-type matching; ExtendedDataFig8 caption; methods: p.40, Benchmarking on the CAGI5 MPRA challenge; paper: Extended Data Fig.8c; Fig.5j; tables: Suppl Table 4 Variant performan; evaluation index 10; sheet rows 12At a glance
Explanatory profile: limited source coverage · Automated source review, 2026-09-17. Review applies to the cited claims; unresolved fields are listed below. Numerical results retain their own review status.
| Property | Description and evidence |
|---|---|
| Dataset and biological context | CAGI5 training/test sets per locus, using native hg19 sequence and GENCODEv19.AlphaGenome Nature 2026 supplementary comparison tables; AlphaGenome Nature 2026 supplementary methods; alphagenome: Journal full-text XML · 'Suppl Table 4 Variant performan'!A12:P12; Supplementary Methods p.30, Chromosome Splits for Variant Benchmarks; Supplementary Methods p.40, Benchmarking on CAGI5 MPRA; Table10 cell-type matching; ExtendedDataFig8 caption; methods: p.40, Benchmarking on the CAGI5 MPRA challenge; paper: Extended Data Fig.8c; Fig.5j; tables: Suppl Table 4 Variant performan; evaluation index 10; sheet rows 12 |
| Split | Fit feature scaling on challenge training data only; tune a separate LassoCV per locus with10-fold cross-validation, then evaluate the challenge test set.AlphaGenome Nature 2026 supplementary comparison tables; AlphaGenome Nature 2026 supplementary methods; alphagenome: Journal full-text XML · 'Suppl Table 4 Variant performan'!A12:P12; Supplementary Methods p.30, Chromosome Splits for Variant Benchmarks; Supplementary Methods p.40, Benchmarking on CAGI5 MPRA; Table10 cell-type matching; ExtendedDataFig8 caption; methods: p.40, Benchmarking on the CAGI5 MPRA challenge; paper: Extended Data Fig.8c; Fig.5j; tables: Suppl Table 4 Variant performan; evaluation index 10; sheet rows 12 |
| Allowed inputs and adaptation | DNase, histone ChIP-seq and target-gene RNA variant-score features across available cell types.AlphaGenome Nature 2026 supplementary comparison tables; AlphaGenome Nature 2026 supplementary methods; alphagenome: Journal full-text XML · 'Suppl Table 4 Variant performan'!A12:P12; Supplementary Methods p.30, Chromosome Splits for Variant Benchmarks; Supplementary Methods p.40, Benchmarking on CAGI5 MPRA; Table10 cell-type matching; ExtendedDataFig8 caption; methods: p.40, Benchmarking on the CAGI5 MPRA challenge; paper: Extended Data Fig.8c; Fig.5j; tables: Suppl Table 4 Variant performan; evaluation index 10; sheet rows 12 |
| Metrics as reported | Lasso pearsonr (DNASE + RNA + CHIP)AlphaGenome Nature 2026 supplementary comparison tables; AlphaGenome Nature 2026 supplementary methods; alphagenome: Journal full-text XML · 'Suppl Table 4 Variant performan'!A12:P12; Supplementary Methods p.30, Chromosome Splits for Variant Benchmarks; Supplementary Methods p.40, Benchmarking on CAGI5 MPRA; Table10 cell-type matching; ExtendedDataFig8 caption; methods: p.40, Benchmarking on the CAGI5 MPRA challenge; paper: Extended Data Fig.8c; Fig.5j; tables: Suppl Table 4 Variant performan; evaluation index 10; sheet rows 12 |
| Aggregation | Mean Pearson correlation across included locus/context tests. The figure-average protocol excludes MYC for LASSO non-convergence and TERT-GBM for matched comparisons.AlphaGenome Nature 2026 supplementary comparison tables; AlphaGenome Nature 2026 supplementary methods; alphagenome: Journal full-text XML · 'Suppl Table 4 Variant performan'!A12:P12; Supplementary Methods p.30, Chromosome Splits for Variant Benchmarks; Supplementary Methods p.40, Benchmarking on CAGI5 MPRA; Table10 cell-type matching; ExtendedDataFig8 caption; methods: p.40, Benchmarking on the CAGI5 MPRA challenge; paper: Extended Data Fig.8c; Fig.5j; tables: Suppl Table 4 Variant performan; evaluation index 10; sheet rows 12 |
| Uncertainty | Not reported for these summary-table scores. · Not reported in inspected sourcesAlphaGenome Nature 2026 supplementary comparison tables; AlphaGenome Nature 2026 supplementary methods; alphagenome: Journal full-text XML · 'Suppl Table 4 Variant performan'!A12:P12; Supplementary Methods p.30, Chromosome Splits for Variant Benchmarks; Supplementary Methods p.40, Benchmarking on CAGI5 MPRA; Table10 cell-type matching; ExtendedDataFig8 caption; methods: p.40, Benchmarking on the CAGI5 MPRA challenge; paper: Extended Data Fig.8c; Fig.5j; tables: Suppl Table 4 Variant performan; evaluation index 10; sheet rows 12 |
| Entity type | protocol |
| Organisms | Not extracted or verified for this record. |
| Assays | Not extracted or verified for this record. |
| Baselines | Not extracted or verified for this record. |
How it works
Conceptual summary of the cited procedure; model-specific conditions are given below.
Read the diagram as text
- Compute multimodal CAGI5 features
- Scale from training data
- Fit per-locus cross-validated LASSO
- Evaluate test-locus correlations
What is tested
Does supervised combination of sequence-derived modalities better predict CAGI5 reporter-assay effects?
AlphaGenome Nature 2026 supplementary comparison tables; AlphaGenome Nature 2026 supplementary methods; alphagenome: Journal full-text XML · 'Suppl Table 4 Variant performan'!A12:P12; Supplementary Methods p.30, Chromosome Splits for Variant Benchmarks; Supplementary Methods p.40, Benchmarking on CAGI5 MPRA; Table10 cell-type matching; ExtendedDataFig8 caption; methods: p.40, Benchmarking on the CAGI5 MPRA challenge; paper: Extended Data Fig.8c; Fig.5j; tables: Suppl Table 4 Variant performan; evaluation index 10; sheet rows 12Procedure
Generate multimodal features using the Borzoi-comparison scoring strategy, standardize from training statistics and fit per-locus LASSO.
AlphaGenome Nature 2026 supplementary comparison tables; AlphaGenome Nature 2026 supplementary methods; alphagenome: Journal full-text XML · 'Suppl Table 4 Variant performan'!A12:P12; Supplementary Methods p.30, Chromosome Splits for Variant Benchmarks; Supplementary Methods p.40, Benchmarking on CAGI5 MPRA; Table10 cell-type matching; ExtendedDataFig8 caption; methods: p.40, Benchmarking on the CAGI5 MPRA challenge; paper: Extended Data Fig.8c; Fig.5j; tables: Suppl Table 4 Variant performan; evaluation index 10; sheet rows 12Tested models and results
Release 2026-09-17-a757f4af4277 · 2 evaluations · 2 metric rows. Different protocols are not a single leaderboard.
| Metric and finding | Coverage and uncertainty | Evidence |
|---|---|---|
| borzoi-ensemble: multimodal features + LASSO (paper Table 4): Supervised multimodal CAGI5 MPRA prediction Model: borzoi-ensemble: multimodal features + LASSO (paper Table 4) · Benchmark: Supervised multimodal CAGI5 MPRA prediction (AlphaGenome paper) · Dataset: Supervised multimodal CAGI5 MPRA prediction: evaluated data subset Generate multimodal features using the Borzoi-comparison scoring strategy, standardize from training statistics and fit per-locus LASSO. Author-reported evaluation · Evaluation metadata: needs review | ||
| 0.615 Lasso pearsonr (DNASE + RNA + CHIP) Unit: correlation · Direction: higher Aggregation: Mean Pearson correlation across included locus/context tests. The figure-average protocol excludes MYC for LASSO non-convergence and TERT-GBM for matched comparisons. | Uncertainty: unreported Scored: Not reported · Eligible: Not reported | source checkedAlphaGenome Nature 2026 supplementary comparison tables · 'Suppl Table 4 Variant performan'!L12 Source checking is not independent reproduction. |
| AlphaGenome distilled multimodal DNase, RNA-seq and histone ChIP-seq features → per-locus LassoCV: Supervised multimodal CAGI5 MPRA prediction Model: AlphaGenome distilled multimodal DNase, RNA-seq and histone ChIP-seq features → per-locus LassoCV · Benchmark: Supervised multimodal CAGI5 MPRA prediction (AlphaGenome paper) · Dataset: Supervised multimodal CAGI5 MPRA prediction: evaluated data subset Generate multimodal features using the Borzoi-comparison scoring strategy, standardize from training statistics and fit per-locus LASSO. Author-reported evaluation · Evaluation metadata: needs review | ||
| 0.65 Lasso pearsonr (DNASE + RNA + CHIP) Unit: correlation · Direction: higher Aggregation: Mean Pearson correlation across included locus/context tests. The figure-average protocol excludes MYC for LASSO non-convergence and TERT-GBM for matched comparisons. | Uncertainty: unreported Scored: Not reported · Eligible: Not reported | source checkedAlphaGenome Nature 2026 supplementary comparison tables · 'Suppl Table 4 Variant performan'!M12 Source checking is not independent reproduction. |
Strengths and limitations
Limitations and conditions
- This is supervised pipeline performance, not a zero-shot base-model result. The F9 exclusion described for DNase-only LASSO must not be automatically applied to this multimodal row.AlphaGenome Nature 2026 supplementary comparison tables; AlphaGenome Nature 2026 supplementary methods; alphagenome: Journal full-text XML · 'Suppl Table 4 Variant performan'!A12:P12; Supplementary Methods p.30, Chromosome Splits for Variant Benchmarks; Supplementary Methods p.40, Benchmarking on CAGI5 MPRA; Table10 cell-type matching; ExtendedDataFig8 caption; methods: p.40, Benchmarking on the CAGI5 MPRA challenge; paper: Extended Data Fig.8c; Fig.5j; tables: Suppl Table 4 Variant performan; evaluation index 10; sheet rows 12
- This is an author-reported protocol, not a rewire rerun. Different datasets, processing and adaptations cannot support an unrestricted leaderboard.AlphaGenome Nature 2026 supplementary comparison tables; AlphaGenome Nature 2026 supplementary methods; alphagenome: Journal full-text XML · 'Suppl Table 4 Variant performan'!A12:P12; Supplementary Methods p.30, Chromosome Splits for Variant Benchmarks; Supplementary Methods p.40, Benchmarking on CAGI5 MPRA; Table10 cell-type matching; ExtendedDataFig8 caption; methods: p.40, Benchmarking on the CAGI5 MPRA challenge; paper: Extended Data Fig.8c; Fig.5j; tables: Suppl Table 4 Variant performan; evaluation index 10; sheet rows 12
Profile review details
Primary-source transcription and separate automated source review. No human sign-off or independent experimental reproduction.
Stable record: alphagenome-2026-t4-protocol-10Evidence table
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.
42 evidence rows matching the loaded filters
| Property and statement | Original source and location | Review and provenance |
|---|---|---|
| Diagram caption Conceptual summary of the cited procedure; model-specific conditions are given below. Individual claims | alphagenome: Journal full-text XML 'Suppl Table 4 Variant performan'!A12:P12; Supplementary Methods p.30, Chromosome Splits for Variant Benchmarks; Supplementary Methods p.40, Benchmarking on CAGI5 MPRA; Table10 cell-type matching; ExtendedDataFig8 caption; methods: p.40, Benchmarking on the CAGI5 MPRA challenge; paper: Extended Data Fig.8c; Fig.5j; tables: Suppl Table 4 Variant performan; evaluation index 10; sheet rows 12 Shared locator for this statement’s cited sources; not a separate locator for each citation. Version: Retrieved page snapshot; no immutable publisher revision supplied | source checked automated source review · 2026-09-17 Audit detailsPrimary-source transcription and separate automated source review. No human sign-off or independent experimental reproduction. Field: Source artifact SHA-256: Hash scope: SHA-256 of retrieved original artifact bytes Format: original_artifact |
| Diagram caption Conceptual summary of the cited procedure; model-specific conditions are given below. Individual claims | AlphaGenome Nature 2026 supplementary methods 'Suppl Table 4 Variant performan'!A12:P12; Supplementary Methods p.30, Chromosome Splits for Variant Benchmarks; Supplementary Methods p.40, Benchmarking on CAGI5 MPRA; Table10 cell-type matching; ExtendedDataFig8 caption; methods: p.40, Benchmarking on the CAGI5 MPRA challenge; paper: Extended Data Fig.8c; Fig.5j; tables: Suppl Table 4 Variant performan; evaluation index 10; sheet rows 12 Shared locator for this statement’s cited sources; not a separate locator for each citation. Version: Supplement to Nature version of record, 28 January 2026; content hash pinned | source checked automated source review · 2026-09-17 Audit detailsPrimary-source transcription and separate automated source review. No human sign-off or independent experimental reproduction. Field: Source artifact SHA-256: Hash scope: SHA-256 of exact publisher PDF bytes Format: original_pdf |
| Diagram caption Conceptual summary of the cited procedure; model-specific conditions are given below. Individual claims | AlphaGenome Nature 2026 supplementary comparison tables 'Suppl Table 4 Variant performan'!A12:P12; Supplementary Methods p.30, Chromosome Splits for Variant Benchmarks; Supplementary Methods p.40, Benchmarking on CAGI5 MPRA; Table10 cell-type matching; ExtendedDataFig8 caption; methods: p.40, Benchmarking on the CAGI5 MPRA challenge; paper: Extended Data Fig.8c; Fig.5j; tables: Suppl Table 4 Variant performan; evaluation index 10; sheet rows 12 Shared locator for this statement’s cited sources; not a separate locator for each citation. Version: Nature version of record, 28 January 2026 | source checked automated source review · 2026-09-17 Audit detailsPrimary-source transcription and separate automated source review. No human sign-off or independent experimental reproduction. Field: Source artifact SHA-256: Hash scope: SHA-256 of exact retrieved original artifact bytes Format: xlsx |
| Diagram steps ["Compute multimodal CAGI5 features","Scale from training data","Fit per-locus cross-validated LASSO","Evaluate test-locus correlations"] Individual claims | alphagenome: Journal full-text XML 'Suppl Table 4 Variant performan'!A12:P12; Supplementary Methods p.30, Chromosome Splits for Variant Benchmarks; Supplementary Methods p.40, Benchmarking on CAGI5 MPRA; Table10 cell-type matching; ExtendedDataFig8 caption; methods: p.40, Benchmarking on the CAGI5 MPRA challenge; paper: Extended Data Fig.8c; Fig.5j; tables: Suppl Table 4 Variant performan; evaluation index 10; sheet rows 12 Shared locator for this statement’s cited sources; not a separate locator for each citation. Version: Retrieved page snapshot; no immutable publisher revision supplied | source checked automated source review · 2026-09-17 Audit detailsPrimary-source transcription and separate automated source review. No human sign-off or independent experimental reproduction. Field: Source artifact SHA-256: Hash scope: SHA-256 of retrieved original artifact bytes Format: original_artifact |
| Diagram steps ["Compute multimodal CAGI5 features","Scale from training data","Fit per-locus cross-validated LASSO","Evaluate test-locus correlations"] Individual claims | AlphaGenome Nature 2026 supplementary methods 'Suppl Table 4 Variant performan'!A12:P12; Supplementary Methods p.30, Chromosome Splits for Variant Benchmarks; Supplementary Methods p.40, Benchmarking on CAGI5 MPRA; Table10 cell-type matching; ExtendedDataFig8 caption; methods: p.40, Benchmarking on the CAGI5 MPRA challenge; paper: Extended Data Fig.8c; Fig.5j; tables: Suppl Table 4 Variant performan; evaluation index 10; sheet rows 12 Shared locator for this statement’s cited sources; not a separate locator for each citation. Version: Supplement to Nature version of record, 28 January 2026; content hash pinned | source checked automated source review · 2026-09-17 Audit detailsPrimary-source transcription and separate automated source review. No human sign-off or independent experimental reproduction. Field: Source artifact SHA-256: Hash scope: SHA-256 of exact publisher PDF bytes Format: original_pdf |
| Diagram steps ["Compute multimodal CAGI5 features","Scale from training data","Fit per-locus cross-validated LASSO","Evaluate test-locus correlations"] Individual claims | AlphaGenome Nature 2026 supplementary comparison tables 'Suppl Table 4 Variant performan'!A12:P12; Supplementary Methods p.30, Chromosome Splits for Variant Benchmarks; Supplementary Methods p.40, Benchmarking on CAGI5 MPRA; Table10 cell-type matching; ExtendedDataFig8 caption; methods: p.40, Benchmarking on the CAGI5 MPRA challenge; paper: Extended Data Fig.8c; Fig.5j; tables: Suppl Table 4 Variant performan; evaluation index 10; sheet rows 12 Shared locator for this statement’s cited sources; not a separate locator for each citation. Version: Nature version of record, 28 January 2026 | source checked automated source review · 2026-09-17 Audit detailsPrimary-source transcription and separate automated source review. No human sign-off or independent experimental reproduction. Field: Source artifact SHA-256: Hash scope: SHA-256 of exact retrieved original artifact bytes Format: xlsx |
| Diagram title Supervised multimodal CAGI5 MPRA prediction: evaluation procedure Individual claims | alphagenome: Journal full-text XML 'Suppl Table 4 Variant performan'!A12:P12; Supplementary Methods p.30, Chromosome Splits for Variant Benchmarks; Supplementary Methods p.40, Benchmarking on CAGI5 MPRA; Table10 cell-type matching; ExtendedDataFig8 caption; methods: p.40, Benchmarking on the CAGI5 MPRA challenge; paper: Extended Data Fig.8c; Fig.5j; tables: Suppl Table 4 Variant performan; evaluation index 10; sheet rows 12 Shared locator for this statement’s cited sources; not a separate locator for each citation. Version: Retrieved page snapshot; no immutable publisher revision supplied | source checked automated source review · 2026-09-17 Audit detailsPrimary-source transcription and separate automated source review. No human sign-off or independent experimental reproduction. Field: Source artifact SHA-256: Hash scope: SHA-256 of retrieved original artifact bytes Format: original_artifact |
| Diagram title Supervised multimodal CAGI5 MPRA prediction: evaluation procedure Individual claims | AlphaGenome Nature 2026 supplementary methods 'Suppl Table 4 Variant performan'!A12:P12; Supplementary Methods p.30, Chromosome Splits for Variant Benchmarks; Supplementary Methods p.40, Benchmarking on CAGI5 MPRA; Table10 cell-type matching; ExtendedDataFig8 caption; methods: p.40, Benchmarking on the CAGI5 MPRA challenge; paper: Extended Data Fig.8c; Fig.5j; tables: Suppl Table 4 Variant performan; evaluation index 10; sheet rows 12 Shared locator for this statement’s cited sources; not a separate locator for each citation. Version: Supplement to Nature version of record, 28 January 2026; content hash pinned | source checked automated source review · 2026-09-17 Audit detailsPrimary-source transcription and separate automated source review. No human sign-off or independent experimental reproduction. Field: Source artifact SHA-256: Hash scope: SHA-256 of exact publisher PDF bytes Format: original_pdf |
| Diagram title Supervised multimodal CAGI5 MPRA prediction: evaluation procedure Individual claims | AlphaGenome Nature 2026 supplementary comparison tables 'Suppl Table 4 Variant performan'!A12:P12; Supplementary Methods p.30, Chromosome Splits for Variant Benchmarks; Supplementary Methods p.40, Benchmarking on CAGI5 MPRA; Table10 cell-type matching; ExtendedDataFig8 caption; methods: p.40, Benchmarking on the CAGI5 MPRA challenge; paper: Extended Data Fig.8c; Fig.5j; tables: Suppl Table 4 Variant performan; evaluation index 10; sheet rows 12 Shared locator for this statement’s cited sources; not a separate locator for each citation. Version: Nature version of record, 28 January 2026 | source checked automated source review · 2026-09-17 Audit detailsPrimary-source transcription and separate automated source review. No human sign-off or independent experimental reproduction. Field: Source artifact SHA-256: Hash scope: SHA-256 of exact retrieved original artifact bytes Format: xlsx |
| Dataset and biological context CAGI5 training/test sets per locus, using native hg19 sequence and GENCODEv19. Individual claims | alphagenome: Journal full-text XML 'Suppl Table 4 Variant performan'!A12:P12; Supplementary Methods p.30, Chromosome Splits for Variant Benchmarks; Supplementary Methods p.40, Benchmarking on CAGI5 MPRA; Table10 cell-type matching; ExtendedDataFig8 caption; methods: p.40, Benchmarking on the CAGI5 MPRA challenge; paper: Extended Data Fig.8c; Fig.5j; tables: Suppl Table 4 Variant performan; evaluation index 10; sheet rows 12 Shared locator for this statement’s cited sources; not a separate locator for each citation. Version: Retrieved page snapshot; no immutable publisher revision supplied | source checked automated source review · 2026-09-17 Audit detailsPrimary-source transcription and separate automated source review. No human sign-off or independent experimental reproduction. Field: Source artifact SHA-256: Hash scope: SHA-256 of retrieved original artifact bytes Format: original_artifact |
Sources and history
Release 2026-09-17-a757f4af4277 · Record review: needs review
- AlphaGenome Nature 2026 supplementary comparison tables · Original source · Nature version of record, 28 January 2026
- AlphaGenome Nature 2026 supplementary methods · Original source · Supplement to Nature version of record, 28 January 2026; content hash pinned
- alphagenome: Journal full-text XML · Original source · Retrieved page snapshot; no immutable publisher revision supplied
Technical metadata and extraction receipts
Stable ID: alphagenome-2026-t4-protocol-10
- areas
- dna-genomes
- entity level
- protocol
- version
- Nature version of record, 28 January 2026
- source evaluation index
- 10
- source table
- 4
- reference levels
- metric: Lasso pearsonr (DNASE + RNA + CHIP); printed value: 0; numeric value: 0; source locator: Suppl Table 4 Variant performan!J12; note: Source reference quantity for relative-performance calculation, not a measured baseline run.
Related records
- benchmark: borzoi-ensemble: multimodal features + LASSO (paper Table 4): Supervised multimodal CAGI5 MPRA prediction
- benchmark: AlphaGenome distilled multimodal DNase, RNA-seq and histone ChIP-seq features → per-locus LassoCV: Supervised multimodal CAGI5 MPRA prediction
- benchmark: Supervised multimodal CAGI5 MPRA prediction: evaluated data subset