Task

RNA compound-binding site prediction

RNA compound-binding prediction is evaluated at nucleotide level with independent test collections and a structural-generalization check.

Sources

CoBRA: compound binding site prediction using RNA language model · Methods: Dataset preparation; Evaluation metrics; Results: Structure-based split dataset evaluation; cached text lines 9–13, 35–36, 61–65

1 evaluation · 1 metric row

At a glance

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.

Data, procedure and scoring
Property	Description and evidence
Datasets	HARIBOSS and TR60 training resources, with RB9, JL10, TL12 and TE18 reserved for testing. Sources CoBRA: compound binding site prediction using RNA language model · Methods: Dataset preparation; Evaluation metrics; Results: Structure-based split dataset evaluation; cached text lines 9–13, 35–36, 61–65
Splits	Combined HARIBOSS/TR60 data use an 80:10:10 internal partition; RB9, JL10, TL12 and TE18 remain external tests. RNABind structural splits define an additional retraining setting. Sources CoBRA: compound binding site prediction using RNA language model · Methods: Dataset preparation; Evaluation metrics; Results: Structure-based split dataset evaluation; cached text lines 9–13, 35–36, 61–65
Metrics	Precision, recall, F1, MCC, AUROC and AUPRC. Sources CoBRA: compound binding site prediction using RNA language model · Methods: Dataset preparation; Evaluation metrics; Results: Structure-based split dataset evaluation; cached text lines 9–13, 35–36, 61–65
Baselines	Comparisons include RNABind and other structure-aware methods in the structural-split evaluation. Sources CoBRA: compound binding site prediction using RNA language model · Methods: Dataset preparation; Evaluation metrics; Results: Structure-based split dataset evaluation; cached text lines 9–13, 35–36, 61–65
Leakage controls	Sequences overlapping training and the four external tests are removed; structural-split retraining separately tests structure-level generalization. Sources CoBRA: compound binding site prediction using RNA language model · Methods: Dataset preparation; Evaluation metrics; Results: Structure-based split dataset evaluation; cached text lines 9–13, 35–36, 61–65
Uncertainty	The cited text-accessible evaluation sections give no confidence-interval, resampling or repeat-run error-bar specification. Image-only tables and uninspected supplements are outside this absence claim. · Not reported in inspected sources Sources CoBRA: compound binding site prediction using RNA language model · Methods: Dataset preparation; Evaluation metrics; Results: Structure-based split dataset evaluation; cached text lines 9–13, 35–36, 61–65
Entity type	Paper-specific computational evaluation protocol. Sources CoBRA: compound binding site prediction using RNA language model · Methods: Dataset preparation; Evaluation metrics; Results: Structure-based split dataset evaluation; cached text lines 9–13, 35–36, 61–65
Organisms	The evaluation pools PDB-derived RNA–ligand complexes from HARIBOSS, TR60, RB9, TL12, JL10 and TE18. Dataset preparation filters ligands, chain length and overlapping sequences, but does not enumerate the organism composition of each resulting subset. · Not reported in inspected sources Sources CoBRA: compound binding site prediction using RNA language model · Material and methods: Dataset preparation
Assays	RNA compound-binding-site annotations. Sources CoBRA: compound binding site prediction using RNA language model · Methods: Dataset preparation; Evaluation metrics; Results: Structure-based split dataset evaluation; cached text lines 9–13, 35–36, 61–65
Allowed inputs	RNA sequence only; explicit RNA structures are not required for CoBRA inference. Sources CoBRA: compound binding site prediction using RNA language model · Methods: Dataset preparation; Evaluation metrics; Results: Structure-based split dataset evaluation; cached text lines 9–13, 35–36, 61–65
Adaptation	Frozen RNA-model residue embeddings feed a supervised MLP; language-model parameters are not updated. Sources CoBRA: compound binding site prediction using RNA language model · Methods: Dataset preparation; Evaluation metrics; Results: Structure-based split dataset evaluation; cached text lines 9–13, 35–36, 61–65

How it works

How it worksComputational evaluation flow

Conceptual summary of the cited evaluation; exact task configuration and source version remain part of the protocol.

Sources

Evaluation methodology

HARIBOSS and TR60 training resources, with RB9, JL10, TL12 and TE18 reserved for testing. Combined HARIBOSS/TR60 data use an 80:10:10 internal partition; RB9, JL10, TL12 and TE18 remain external tests. RNABind structural splits define an additional retraining setting. Precision, recall, F1, MCC, AUROC and AUPRC. Comparisons include RNABind and other structure-aware methods in the structural-split evaluation. Sequences overlapping training and the four external tests are removed; structural-split retraining separately tests structure-level generalization.

Sources

Recorded evaluations

Each evaluation records what was tested and under which conditions.

ERNIE-RNA + CoBRA: RNA compound-binding site prediction

Tested entities and results

Release 2026-09-17-d277315f7d76 · 1 evaluation · 1 metric row. Different protocols are not a single leaderboard.

Evaluation setupMetricEvidence origin

Results grouped by the exact reported evaluation
Metric and finding	Coverage and uncertainty	Evidence
ERNIE-RNA + CoBRA: RNA compound-binding site prediction Pipeline: ERNIE-RNA + CoBRATask: RNA compound-binding site predictionDataset: CoBRA compound-binding test set ERNIE-RNA embedding with TCL focal loss Author-reported evaluation · Evaluation metadata: needs review
0.657 MCC Unit: unitless · Direction: unknown	Uncertainty: not reported in legacy extract Scored: Not reported · Eligible: Not reported	source checkedCoBRA: compound binding site prediction using RNA language model · Table 2, ERNIE-RNA / TCL focal row, MCC column Source checking is not 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
CoBRA: compound binding site prediction using RNA language model	journal full text in PMC	Read source DOI: 10.1093/bib/bbaf713

What is still missing

Baselines in Table 4 are taken from Gao et al., Zhu et al. and Chen et al.; this is a literature comparison, not uniformly rerun methods.
Do not assign the selected CoBRA pipeline score to RiNALMo alone; absent values are unavailable, not zero.

Search and extraction details

primary comparison table screened

Searches

"COBRA" "RNA" "PMC12790621"

Evidence locations

Table 4 footnote
Results: Comparison with other existing methods on benchmark sets

Strengths and limitations

Strengths and considerations

Sequence-overlap removal protects the external tests; a separate structural-split experiment probes a different generalization boundary.
Sources
CoBRA: compound binding site prediction using RNA language model · Methods: Dataset preparation; Evaluation metrics; Results: Structure-based split dataset evaluation; cached text lines 9–13, 35–36, 61–65

Limitations and conditions

Metal and nonmetal compounds are both present. Sequence-only and structure-aware comparators use different input information.
Sources
CoBRA: compound binding site prediction using RNA language model · Methods: Dataset preparation; Evaluation metrics; Results: Structure-based split dataset evaluation; cached text lines 9–13, 35–36, 61–65

Profile review details

Targeted full-paper and supplement review of the outstanding task fields, with original dataset metadata checked where accessible. Source-scoped omissions are explicit; no independent benchmark reproduction or numerical-result change.

Stable record: reported-task-3a3bff34cce634

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.

18 evidence rows matching the loaded filters

Claims, original sources and review scope · Release 2026-09-17-d277315f7d76
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	CoBRA: compound binding site prediction using RNA language model Original source ↗ Methods: Dataset preparation; Evaluation metrics; Results: Structure-based split dataset evaluation; cached text lines 9–13, 35–36, 61–65 Version: journal full text in PMC Retrieved: 2026-09-16T10:38:57.558197+00:00	source checked automated source review · 2026-09-16 Audit details Targeted full-paper and supplement review of the outstanding task fields, with original dataset metadata checked where accessible. Source-scoped omissions are explicit; no independent benchmark reproduction or numerical-result change. Field: `attributes.profile.diagram.caption` Source artifact SHA-256: `8c6a6f00f5fa5f62acf301a66e9e6fa9ef11c7a05ad9b7447d2ade2ce8eba793` Hash scope: Hash scope not separately documented; inspect source record Inspected artifact
Diagram steps ["Input: RNA sequence only; explicit RNA structures are not required for CoBRA inference.","Evaluation: Combined HARIBOSS/TR60 data use an 80:10:10 internal partition; RB9, JL10, TL12 and TE18 remain external tests. RNABind structural splits define an additional retraining setting.","Readout: Precision, recall, F1, MCC, AUROC and AUPRC."] Individual claims	CoBRA: compound binding site prediction using RNA language model Original source ↗ Methods: Dataset preparation; Evaluation metrics; Results: Structure-based split dataset evaluation; cached text lines 9–13, 35–36, 61–65 Version: journal full text in PMC Retrieved: 2026-09-16T10:38:57.558197+00:00	source checked automated source review · 2026-09-16 Audit details Targeted full-paper and supplement review of the outstanding task fields, with original dataset metadata checked where accessible. Source-scoped omissions are explicit; no independent benchmark reproduction or numerical-result change. Field: `attributes.profile.diagram.steps` Source artifact SHA-256: `8c6a6f00f5fa5f62acf301a66e9e6fa9ef11c7a05ad9b7447d2ade2ce8eba793` Hash scope: Hash scope not separately documented; inspect source record Inspected artifact
Diagram title Computational evaluation flow Individual claims	CoBRA: compound binding site prediction using RNA language model Original source ↗ Methods: Dataset preparation; Evaluation metrics; Results: Structure-based split dataset evaluation; cached text lines 9–13, 35–36, 61–65 Version: journal full text in PMC Retrieved: 2026-09-16T10:38:57.558197+00:00	source checked automated source review · 2026-09-16 Audit details Targeted full-paper and supplement review of the outstanding task fields, with original dataset metadata checked where accessible. Source-scoped omissions are explicit; no independent benchmark reproduction or numerical-result change. Field: `attributes.profile.diagram.title` Source artifact SHA-256: `8c6a6f00f5fa5f62acf301a66e9e6fa9ef11c7a05ad9b7447d2ade2ce8eba793` Hash scope: Hash scope not separately documented; inspect source record Inspected artifact
Datasets HARIBOSS and TR60 training resources, with RB9, JL10, TL12 and TE18 reserved for testing. Individual claims	CoBRA: compound binding site prediction using RNA language model Original source ↗ Methods: Dataset preparation; Evaluation metrics; Results: Structure-based split dataset evaluation; cached text lines 9–13, 35–36, 61–65 Version: journal full text in PMC Retrieved: 2026-09-16T10:38:57.558197+00:00	source checked automated source review · 2026-09-16 Audit details Targeted full-paper and supplement review of the outstanding task fields, with original dataset metadata checked where accessible. Source-scoped omissions are explicit; no independent benchmark reproduction or numerical-result change. Field: `attributes.profile.facts.0.value` Source artifact SHA-256: `8c6a6f00f5fa5f62acf301a66e9e6fa9ef11c7a05ad9b7447d2ade2ce8eba793` Hash scope: Hash scope not separately documented; inspect source record Inspected artifact
Splits Combined HARIBOSS/TR60 data use an 80:10:10 internal partition; RB9, JL10, TL12 and TE18 remain external tests. RNABind structural splits define an additional retraining setting. Individual claims	CoBRA: compound binding site prediction using RNA language model Original source ↗ Methods: Dataset preparation; Evaluation metrics; Results: Structure-based split dataset evaluation; cached text lines 9–13, 35–36, 61–65 Version: journal full text in PMC Retrieved: 2026-09-16T10:38:57.558197+00:00	source checked automated source review · 2026-09-16 Audit details Targeted full-paper and supplement review of the outstanding task fields, with original dataset metadata checked where accessible. Source-scoped omissions are explicit; no independent benchmark reproduction or numerical-result change. Field: `attributes.profile.facts.1.value` Source artifact SHA-256: `8c6a6f00f5fa5f62acf301a66e9e6fa9ef11c7a05ad9b7447d2ade2ce8eba793` Hash scope: Hash scope not separately documented; inspect source record Inspected artifact
Adaptation Frozen RNA-model residue embeddings feed a supervised MLP; language-model parameters are not updated. Individual claims	CoBRA: compound binding site prediction using RNA language model Original source ↗ Methods: Dataset preparation; Evaluation metrics; Results: Structure-based split dataset evaluation; cached text lines 9–13, 35–36, 61–65 Version: journal full text in PMC Retrieved: 2026-09-16T10:38:57.558197+00:00	source checked automated source review · 2026-09-16 Audit details Targeted full-paper and supplement review of the outstanding task fields, with original dataset metadata checked where accessible. Source-scoped omissions are explicit; no independent benchmark reproduction or numerical-result change. Field: `attributes.profile.facts.10.value` Source artifact SHA-256: `8c6a6f00f5fa5f62acf301a66e9e6fa9ef11c7a05ad9b7447d2ade2ce8eba793` Hash scope: Hash scope not separately documented; inspect source record Inspected artifact
Metrics Precision, recall, F1, MCC, AUROC and AUPRC. Individual claims	CoBRA: compound binding site prediction using RNA language model Original source ↗ Methods: Dataset preparation; Evaluation metrics; Results: Structure-based split dataset evaluation; cached text lines 9–13, 35–36, 61–65 Version: journal full text in PMC Retrieved: 2026-09-16T10:38:57.558197+00:00	source checked automated source review · 2026-09-16 Audit details Targeted full-paper and supplement review of the outstanding task fields, with original dataset metadata checked where accessible. Source-scoped omissions are explicit; no independent benchmark reproduction or numerical-result change. Field: `attributes.profile.facts.2.value` Source artifact SHA-256: `8c6a6f00f5fa5f62acf301a66e9e6fa9ef11c7a05ad9b7447d2ade2ce8eba793` Hash scope: Hash scope not separately documented; inspect source record Inspected artifact
Baselines Comparisons include RNABind and other structure-aware methods in the structural-split evaluation. Individual claims	CoBRA: compound binding site prediction using RNA language model Original source ↗ Methods: Dataset preparation; Evaluation metrics; Results: Structure-based split dataset evaluation; cached text lines 9–13, 35–36, 61–65 Version: journal full text in PMC Retrieved: 2026-09-16T10:38:57.558197+00:00	source checked automated source review · 2026-09-16 Audit details Targeted full-paper and supplement review of the outstanding task fields, with original dataset metadata checked where accessible. Source-scoped omissions are explicit; no independent benchmark reproduction or numerical-result change. Field: `attributes.profile.facts.3.value` Source artifact SHA-256: `8c6a6f00f5fa5f62acf301a66e9e6fa9ef11c7a05ad9b7447d2ade2ce8eba793` Hash scope: Hash scope not separately documented; inspect source record Inspected artifact
Leakage controls Sequences overlapping training and the four external tests are removed; structural-split retraining separately tests structure-level generalization. Individual claims	CoBRA: compound binding site prediction using RNA language model Original source ↗ Methods: Dataset preparation; Evaluation metrics; Results: Structure-based split dataset evaluation; cached text lines 9–13, 35–36, 61–65 Version: journal full text in PMC Retrieved: 2026-09-16T10:38:57.558197+00:00	source checked automated source review · 2026-09-16 Audit details Targeted full-paper and supplement review of the outstanding task fields, with original dataset metadata checked where accessible. Source-scoped omissions are explicit; no independent benchmark reproduction or numerical-result change. Field: `attributes.profile.facts.4.value` Source artifact SHA-256: `8c6a6f00f5fa5f62acf301a66e9e6fa9ef11c7a05ad9b7447d2ade2ce8eba793` Hash scope: Hash scope not separately documented; inspect source record Inspected artifact
Uncertainty The cited text-accessible evaluation sections give no confidence-interval, resampling or repeat-run error-bar specification. Image-only tables and uninspected supplements are outside this absence claim. Individual claims	CoBRA: compound binding site prediction using RNA language model Original source ↗ Methods: Dataset preparation; Evaluation metrics; Results: Structure-based split dataset evaluation; cached text lines 9–13, 35–36, 61–65 Version: journal full text in PMC Retrieved: 2026-09-16T10:38:57.558197+00:00	unreported automated source review · 2026-09-16 Audit details Targeted full-paper and supplement review of the outstanding task fields, with original dataset metadata checked where accessible. Source-scoped omissions are explicit; no independent benchmark reproduction or numerical-result change. Field: `attributes.profile.facts.5.value` Source artifact SHA-256: `8c6a6f00f5fa5f62acf301a66e9e6fa9ef11c7a05ad9b7447d2ade2ce8eba793` Hash scope: Hash scope not separately documented; inspect source record Inspected artifact

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

Sources and history

Release 2026-09-17-d277315f7d76 · Record review: needs review

2 source records and release history

CoBRA: compound binding site prediction using RNA language model · Original source · journal full text in PMC
CoBRA: compound binding site prediction using RNA language model · Original source · journal full text in PMC

Download this release

Technical metadata and extraction receipts

Stable ID: reported-task-3a3bff34cce634

areas: rna-transcriptomes
tasks: RNA compound-binding site prediction
entity level: task
version: Not reported
task: RNA compound-binding 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-cobra-rna-binding-2026; inspected locators: Table 4 footnote; Results: Comparison with other existing methods on benchmark sets; searched queries: "COBRA" "RNA" "PMC12790621"; gaps: Baselines in Table 4 are taken from Gao et al., Zhu et al. and Chen et al.; this is a literature comparison, not uniformly rerun methods.; Do not assign the selected CoBRA pipeline score to RiNALMo alone; absent values are unavailable, not zero.; 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: cobra-rna-binding-2026; source locator: Methods: Dataset preparation; Evaluation metrics; Results: Structure-based split dataset evaluation; cached text lines 9–13, 35–36, 61–65; 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: CoBRA compound-binding test set
benchmark: ERNIE-RNA + CoBRA: RNA compound-binding site prediction

Suggest a correction