Datasets
GTEx liver RNA-seq annotations in Alu-associated regions; labels distinguish editing levels.
Task
A-to-I RNA editing site prediction
An RNA editing-site classifier is compared with sequence-model baselines on held-out human liver annotations.
Sources
ADAR-GPT: A continually fine-tuned language model for predicting A-to-I RNA editing sites · Methods: Dataset Construction and Annotation; Dataset Design and Labeling; Evaluation and Reproducibility; Table 2; cached text lines 91–110Overview
Metrics
Accuracy, precision, recall, specificity and F1; probability-based AUROC and AUPRC.
Allowed inputs
RNA sequence around candidate editing sites.
Sources
ADAR-GPT: A continually fine-tuned language model for predicting A-to-I RNA editing sites · Methods: Dataset Construction and Annotation; Dataset Design and Labeling; Evaluation and Reproducibility; Table 2; cached text lines 91–110Evaluation procedure diagram
How it worksComputational evaluation flow
Conceptual summary of the cited evaluation; exact task configuration and source version remain part of the protocol.
Sources
ADAR-GPT: A continually fine-tuned language model for predicting A-to-I RNA editing sites · Methods: Dataset Construction and Annotation; Dataset Design and Labeling; Evaluation and Reproducibility; Table 2; cached text lines 91–110limited source coverage · Automated source review, 2026-09-16. All specifications and missing details
Results
Results are available, but no reviewed comparison panel is linked in this release.
All evaluations
1 evaluation · 1 result. Different protocols are not a single leaderboard.
Filter evaluations
Applied filters: All linked evaluations
| Tested configuration | Protocol and dataset | Finding | Evidence and details |
|---|---|---|---|
| Configuration: ADAR-GPT continual | Task: A-to-I RNA editing site prediction Dataset: liver editing sites | 0.763 F1 fraction · unknown Uncertainty: Not reported Coverage: Not reported scored / Not reported eligible | Author-reported evaluation · Source checkedMethods, coverage and sourceADAR-GPT continual: A-to-I RNA editing site prediction Curriculum plus 15% fine-tuning; 201-nt sequence windows; decision threshold 0.5 Aggregation: Not reported ADAR-GPT: A continually fine-tuned language model for predicting A-to-I RNA editing sites · Table 2, Adar-GPT (continual) row, F1 column |
Source checking is not independent reproduction. Release 2026-09-29-06401fd5b220.
Methods and evaluation design
Procedure, tasks and evaluated configurations
How it works
Evaluation methodology
GTEx liver RNA-seq annotations in Alu-associated regions; labels distinguish editing levels. Random 80:20 splits within disjoint site groups; final comparisons use the held-out highest-threshold validation group. Accuracy, precision, recall, specificity and F1; probability-based AUROC and AUPRC. Static fine-tuning, pretrained and fine-tuned EditPredict, and fine-tuned RNA-FM. Editing sites are assigned to nonoverlapping groups. Independence of overlapping sequence windows or donors was not established in this review. Five inference repeats quantify prediction variability; they do not measure variability across retraining or independent cohorts.
Sources
ADAR-GPT: A continually fine-tuned language model for predicting A-to-I RNA editing sites · Methods: Dataset Construction and Annotation; Dataset Design and Labeling; Evaluation and Reproducibility; Table 2; cached text lines 91–110Recorded evaluations
Each evaluation records what was tested and under which conditions.
Run instructions
No runnable recipe has been reviewed for this task. Dataset access, model requirements, licences and compute requirements must be checked against its sources before execution.
A task describes a biological question. Choose a linked protocol to obtain concrete split and scoring instructions.
Strengths, limitations and unresolved questions
Strengths and limitations
Strengths and considerations
No source-reviewed explanatory claims are recorded here yet.
Limitations and conditions
- The evaluation partition is called validation in the source and is reused during training monitoring; it should not be relabelled an untouched external test set.
Sources
ADAR-GPT: A continually fine-tuned language model for predicting A-to-I RNA editing sites · Methods: Dataset Construction and Annotation; Dataset Design and Labeling; Evaluation and Reproducibility; Table 2; cached text lines 91–110
Profile review details
Task-specific computational methodology and field context checked in the cited primary-source artifact. Source-backed fields, inapplicable evaluator dimensions and unresolved details are distinguished. Numerical results were not reproduced.
Stable record: reported-task-d635fc6c281a27Specifications
Inputs, training, access and other details
Explanatory profile: limited source coverage · Automated source review, 2026-09-16. Review applies to the cited claims; unresolved fields are listed below. Numerical results retain their own review status.
| Property | Description and evidence |
|---|---|
| Datasets | GTEx liver RNA-seq annotations in Alu-associated regions; labels distinguish editing levels.SourcesADAR-GPT: A continually fine-tuned language model for predicting A-to-I RNA editing sites · Methods: Dataset Construction and Annotation; Dataset Design and Labeling; Evaluation and Reproducibility; Table 2; cached text lines 91–110 |
| Splits | Random 80:20 splits within disjoint site groups; final comparisons use the held-out highest-threshold validation group.SourcesADAR-GPT: A continually fine-tuned language model for predicting A-to-I RNA editing sites · Methods: Dataset Construction and Annotation; Dataset Design and Labeling; Evaluation and Reproducibility; Table 2; cached text lines 91–110 |
| Metrics | Accuracy, precision, recall, specificity and F1; probability-based AUROC and AUPRC.SourcesADAR-GPT: A continually fine-tuned language model for predicting A-to-I RNA editing sites · Methods: Dataset Construction and Annotation; Dataset Design and Labeling; Evaluation and Reproducibility; Table 2; cached text lines 91–110 |
| Baselines | Static fine-tuning, pretrained and fine-tuned EditPredict, and fine-tuned RNA-FM.SourcesADAR-GPT: A continually fine-tuned language model for predicting A-to-I RNA editing sites · Methods: Dataset Construction and Annotation; Dataset Design and Labeling; Evaluation and Reproducibility; Table 2; cached text lines 91–110 |
| Leakage controls | Editing sites are assigned to nonoverlapping groups. Independence of overlapping sequence windows or donors was not established in this review.SourcesADAR-GPT: A continually fine-tuned language model for predicting A-to-I RNA editing sites · Methods: Dataset Construction and Annotation; Dataset Design and Labeling; Evaluation and Reproducibility; Table 2; cached text lines 91–110 |
| Uncertainty | Five inference repeats quantify prediction variability; they do not measure variability across retraining or independent cohorts.SourcesADAR-GPT: A continually fine-tuned language model for predicting A-to-I RNA editing sites · Methods: Dataset Construction and Annotation; Dataset Design and Labeling; Evaluation and Reproducibility; Table 2; cached text lines 91–110 |
| Entity type | Paper-specific computational evaluation protocol.SourcesADAR-GPT: A continually fine-tuned language model for predicting A-to-I RNA editing sites · Methods: Dataset Construction and Annotation; Dataset Design and Labeling; Evaluation and Reproducibility; Table 2; cached text lines 91–110 |
| Organisms | Human liver.SourcesADAR-GPT: A continually fine-tuned language model for predicting A-to-I RNA editing sites · Methods: Dataset Construction and Annotation; Dataset Design and Labeling; Evaluation and Reproducibility; Table 2; cached text lines 91–110 |
| Assays | RNA-seq editing annotations in Alu-associated regions.SourcesADAR-GPT: A continually fine-tuned language model for predicting A-to-I RNA editing sites · Methods: Dataset Construction and Annotation; Dataset Design and Labeling; Evaluation and Reproducibility; Table 2; cached text lines 91–110 |
| Allowed inputs | RNA sequence around candidate editing sites.SourcesADAR-GPT: A continually fine-tuned language model for predicting A-to-I RNA editing sites · Methods: Dataset Construction and Annotation; Dataset Design and Labeling; Evaluation and Reproducibility; Table 2; cached text lines 91–110 |
| Adaptation | Task fine-tuning is compared with pretrained and fine-tuned RNA baselines.SourcesADAR-GPT: A continually fine-tuned language model for predicting A-to-I RNA editing sites · Methods: Dataset Construction and Annotation; Dataset Design and Labeling; Evaluation and Reproducibility; Table 2; cached text lines 91–110 |
Evidence
Source checking verifies the cited claim or transcription. It does not establish independent reproduction.
Papers and result coverage
Last literature check: 2026-09-17. Primary-paper discovery and source inspection. Source-checked results are not independently reproduced experiments.
| Paper or primary resource | Version | Reference |
|---|---|---|
| ADAR-GPT: A continually fine-tuned language model for predicting A-to-I RNA editing sites | version of record | Read source DOI: 10.1073/pnas.2529073123 |
Historical gaps recorded on 2026-09-17
The catalogue now holds 1 result rows for this benchmark. A note below about pending extraction describes the state on 2026-09-17 and may since have been answered by a later batch. The result rows and their sources are the current record.
- This is validation-set reporting, not a cleanly identified untouched test set.
- Table 2 calls it the '15% liver validation set', while methods describe held-out 20% validation; retain wording and resolve data partition before external ranking.
- No uncertainty in the table.
Search and extraction details
primary comparison table screened
Searches
- "PMC12798952"
Evidence locations
- Table 2 and caption
- Methods: Evaluation and Reproducibility
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.
17 evidence rows matching the loaded filters
| Property and statement | Original source and location | Review and provenance |
|---|---|---|
| Diagram caption Conceptual summary of the cited evaluation; exact task configuration and source version remain part of the protocol. Individual claims | ADAR-GPT: A continually fine-tuned language model for predicting A-to-I RNA editing sites Methods: Dataset Construction and Annotation; Dataset Design and Labeling; Evaluation and Reproducibility; Table 2; cached text lines 91–110 Version: version of record | source checked automated source review · 2026-09-16 Audit detailsTask-specific computational methodology and field context checked in the cited primary-source artifact. Source-backed fields, inapplicable evaluator dimensions and unresolved details are distinguished. Numerical results were not reproduced. Field: Source artifact SHA-256: Hash scope: Hash scope not separately documented; inspect source record |
Diagram steps
| ADAR-GPT: A continually fine-tuned language model for predicting A-to-I RNA editing sites Methods: Dataset Construction and Annotation; Dataset Design and Labeling; Evaluation and Reproducibility; Table 2; cached text lines 91–110 Version: version of record | source checked automated source review · 2026-09-16 Audit detailsTask-specific computational methodology and field context checked in the cited primary-source artifact. Source-backed fields, inapplicable evaluator dimensions and unresolved details are distinguished. Numerical results were not reproduced. Field: Source artifact SHA-256: Hash scope: Hash scope not separately documented; inspect source record |
| Diagram title Computational evaluation flow Individual claims | ADAR-GPT: A continually fine-tuned language model for predicting A-to-I RNA editing sites Methods: Dataset Construction and Annotation; Dataset Design and Labeling; Evaluation and Reproducibility; Table 2; cached text lines 91–110 Version: version of record | source checked automated source review · 2026-09-16 Audit detailsTask-specific computational methodology and field context checked in the cited primary-source artifact. Source-backed fields, inapplicable evaluator dimensions and unresolved details are distinguished. Numerical results were not reproduced. Field: Source artifact SHA-256: Hash scope: Hash scope not separately documented; inspect source record |
| Datasets GTEx liver RNA-seq annotations in Alu-associated regions; labels distinguish editing levels. Individual claims | ADAR-GPT: A continually fine-tuned language model for predicting A-to-I RNA editing sites Methods: Dataset Construction and Annotation; Dataset Design and Labeling; Evaluation and Reproducibility; Table 2; cached text lines 91–110 Version: version of record | source checked automated source review · 2026-09-16 Audit detailsTask-specific computational methodology and field context checked in the cited primary-source artifact. Source-backed fields, inapplicable evaluator dimensions and unresolved details are distinguished. Numerical results were not reproduced. Field: Source artifact SHA-256: Hash scope: Hash scope not separately documented; inspect source record |
| Splits Random 80:20 splits within disjoint site groups; final comparisons use the held-out highest-threshold validation group. Individual claims | ADAR-GPT: A continually fine-tuned language model for predicting A-to-I RNA editing sites Methods: Dataset Construction and Annotation; Dataset Design and Labeling; Evaluation and Reproducibility; Table 2; cached text lines 91–110 Version: version of record | source checked automated source review · 2026-09-16 Audit detailsTask-specific computational methodology and field context checked in the cited primary-source artifact. Source-backed fields, inapplicable evaluator dimensions and unresolved details are distinguished. Numerical results were not reproduced. Field: Source artifact SHA-256: Hash scope: Hash scope not separately documented; inspect source record |
| Adaptation Task fine-tuning is compared with pretrained and fine-tuned RNA baselines. Individual claims | ADAR-GPT: A continually fine-tuned language model for predicting A-to-I RNA editing sites Methods: Dataset Construction and Annotation; Dataset Design and Labeling; Evaluation and Reproducibility; Table 2; cached text lines 91–110 Version: version of record | source checked automated source review · 2026-09-16 Audit detailsTask-specific computational methodology and field context checked in the cited primary-source artifact. Source-backed fields, inapplicable evaluator dimensions and unresolved details are distinguished. Numerical results were not reproduced. Field: Source artifact SHA-256: Hash scope: Hash scope not separately documented; inspect source record |
| Metrics Accuracy, precision, recall, specificity and F1; probability-based AUROC and AUPRC. Individual claims | ADAR-GPT: A continually fine-tuned language model for predicting A-to-I RNA editing sites Methods: Dataset Construction and Annotation; Dataset Design and Labeling; Evaluation and Reproducibility; Table 2; cached text lines 91–110 Version: version of record | source checked automated source review · 2026-09-16 Audit detailsTask-specific computational methodology and field context checked in the cited primary-source artifact. Source-backed fields, inapplicable evaluator dimensions and unresolved details are distinguished. Numerical results were not reproduced. Field: Source artifact SHA-256: Hash scope: Hash scope not separately documented; inspect source record |
| Baselines Static fine-tuning, pretrained and fine-tuned EditPredict, and fine-tuned RNA-FM. Individual claims | ADAR-GPT: A continually fine-tuned language model for predicting A-to-I RNA editing sites Methods: Dataset Construction and Annotation; Dataset Design and Labeling; Evaluation and Reproducibility; Table 2; cached text lines 91–110 Version: version of record | source checked automated source review · 2026-09-16 Audit detailsTask-specific computational methodology and field context checked in the cited primary-source artifact. Source-backed fields, inapplicable evaluator dimensions and unresolved details are distinguished. Numerical results were not reproduced. Field: Source artifact SHA-256: Hash scope: Hash scope not separately documented; inspect source record |
| Leakage controls Editing sites are assigned to nonoverlapping groups. Independence of overlapping sequence windows or donors was not established in this review. Individual claims | ADAR-GPT: A continually fine-tuned language model for predicting A-to-I RNA editing sites Methods: Dataset Construction and Annotation; Dataset Design and Labeling; Evaluation and Reproducibility; Table 2; cached text lines 91–110 Version: version of record | source checked automated source review · 2026-09-16 Audit detailsTask-specific computational methodology and field context checked in the cited primary-source artifact. Source-backed fields, inapplicable evaluator dimensions and unresolved details are distinguished. Numerical results were not reproduced. Field: Source artifact SHA-256: Hash scope: Hash scope not separately documented; inspect source record |
| Uncertainty Five inference repeats quantify prediction variability; they do not measure variability across retraining or independent cohorts. Individual claims | ADAR-GPT: A continually fine-tuned language model for predicting A-to-I RNA editing sites Methods: Dataset Construction and Annotation; Dataset Design and Labeling; Evaluation and Reproducibility; Table 2; cached text lines 91–110 Version: version of record | source checked automated source review · 2026-09-16 Audit detailsTask-specific computational methodology and field context checked in the cited primary-source artifact. Source-backed fields, inapplicable evaluator dimensions and unresolved details are distinguished. Numerical results were not reproduced. Field: Source artifact SHA-256: Hash scope: Hash scope not separately documented; inspect source record |
Sources and history
View linked audit checks and correction history
Release 2026-09-29-06401fd5b220 · Record review: needs review
2 source records and release history
- ADAR-GPT: A continually fine-tuned language model for predicting A-to-I RNA editing sites · Original source · version of record
- ADAR-GPT: A continually fine-tuned language model for predicting A-to-I RNA editing sites · Original source · version of record
Technical metadata and extraction receipts
Stable ID: reported-task-d635fc6c281a27
- areas
- rna-transcriptomes
- tasks
- A-to-I RNA editing site prediction
- entity level
- task
- version
- Not reported
- task
- A-to-I RNA editing site prediction
- scope note
- Paper-specific evaluation task; protocol completeness requires further extraction.
- benchmark research
- review date: 2026-09-17; status: primary_comparison_table_screened; primary sources: expansion-p3-adar-gpt-editing-2026; inspected locators: Table 2 and caption; Methods: Evaluation and Reproducibility; searched queries: "PMC12798952"; gaps: This is validation-set reporting, not a cleanly identified untouched test set.; Table 2 calls it the '15% liver validation set', while methods describe held-out 20% validation; retain wording and resolve data partition before external ranking.; No uncertainty in the table.; claim scope: Primary-paper discovery and source inspection. Source-checked results are not independently reproduced experiments.
- historical missing metadata
- protocol version: not_reported_in_legacy_extract
- metadata review scope
- historical_missing_metadata preserves the original discovery state. Current descriptive evidence and missingness are recorded in profile.facts; numerical-result review is separate.
- legacy kinds
- benchmark
- entity classification
- review date: 2026-09-17; rationale: This source-scoped record identifies the biological prediction task and holds its paper context. Preserve the existing task identity; exact split, model adaptation and scoring remain in linked evaluations or separate protocol records.; source ids: adar-gpt-editing-2026; source locator: Methods: Dataset Construction and Annotation; Dataset Design and Labeling; Evaluation and Reproducibility; Table 2; cached text lines 91–110; ambiguities: A paper- or suite-specific task may constrain some inputs or metrics; that alone does not make it interchangeable with a complete versioned protocol. No protocol equivalence is inferred.; Some legacy profile Entity type facts use the generic phrase computational evaluation protocol. That boilerplate is not sufficient to establish a single fixed protocol identity or to merge this task with another protocol record.
Related records
- dataset: liver editing sites
- benchmark: ADAR-GPT continual: A-to-I RNA editing site prediction