rewirebio.iobenchmarks
Pipeline

orthrus-large-6

Linear probe, trained per mRNABench task, on frozen embeddings averaged over nucleotides from orthrus-large-6, an Orthrus checkpoint pretrained with splice and codon tracks; Orthrus is a Mamba mRNA foundation model pretrained by contrastive learning on mRNA splice isoforms and orthologs.

8 evaluations · 8 results

Overview

An explanatory profile has not yet been reviewed for this record.

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

Model Capabilities & Hardware Requirements

Deployment facts recorded for orthrus-large-6. Fields the record does not carry are marked not recorded.

ParametersNot recorded
Minimum VRAMNot recorded
Inference LatencyNot recorded
Context WindowNot recorded
AccessNot recorded
LicenseNot recorded

Evaluations and results

8 evaluations · 8 results. Different protocols are not a single leaderboard.

Sorted by AUPRC (higher is better). The best value in each column is highlighted. Decimals are rounded for display; each value links to the printed value and its source.

All 8 result rows with coverage, uncertainty and sources
Filter evaluations

Applied filters: All linked evaluations

Exact evaluated configurations and original reported results
Tested configurationProtocol and datasetFindingEvidence and details
Pipeline: orthrus-large-6Protocol: mRNABench variant probes eCLIP: eCLIP binding site prediction
Dataset subset: mRNABench eCLIP (mRNABench split)
0.436 ± 0.004 AUPRC
fraction · higher

Uncertainty: 95% CI ± 0.004

Coverage: Not reported scored / Not reported eligible

Author-reported evaluation · Source checked
Methods, coverage and source

orthrus-large-6 on mRNABench variant probes eCLIP: eCLIP binding site prediction

Frozen transcript representations with task-specific linear probes, or named supervised/naive controls. Homology splits where possible; random splits for MRL-MPRA, MRL-HL-Pair and VEP. Reported mean over ten splits; Appendix C enumerates nine seeds. Table captions say ten random splits, while Methods specify homology splits where possible. This source ambiguity and the nine explicitly listed seeds are retained; exact split manifests remain unextracted. Appendix C uses micro-averaged multilabel metrics.

Aggregation: Not reported

mRNABench: A curated benchmark for mature mRNA property and function prediction · Table 5 (XML T2), data row 34, XML tr 36, column 4 (eCLIP | AUPRC), model orthrus-large-6
Pipeline: orthrus-large-6Protocol: mRNABench variant probes GO: Gene Ontology term prediction
Dataset subset: mRNABench GO (mRNABench split)
0.429 ± 0.004 AUPRC
fraction · higher

Uncertainty: 95% CI ± 0.004

Coverage: Not reported scored / Not reported eligible

Author-reported evaluation · Source checked
Methods, coverage and source

orthrus-large-6 on mRNABench variant probes GO: Gene Ontology term prediction

Frozen transcript representations with task-specific linear probes, or named supervised/naive controls. Homology splits where possible; random splits for MRL-MPRA, MRL-HL-Pair and VEP. Reported mean over ten splits; Appendix C enumerates nine seeds. Table captions say ten random splits, while Methods specify homology splits where possible. This source ambiguity and the nine explicitly listed seeds are retained; exact split manifests remain unextracted. Appendix C uses micro-averaged multilabel metrics.

Aggregation: Not reported

mRNABench: A curated benchmark for mature mRNA property and function prediction · Table 6 (XML T3), data row 34, XML tr 36, column 5 (GO | AUPRC), model orthrus-large-6
Pipeline: orthrus-large-6Protocol: mRNABench variant probes HL: mRNA half life
Dataset subset: mRNABench HL (mRNABench split)
0.667 ± 0.005 Pearson
unitless · higher

Uncertainty: 95% CI ± 0.005

Coverage: Not reported scored / Not reported eligible

Author-reported evaluation · Source checked
Methods, coverage and source

orthrus-large-6 on mRNABench variant probes HL: mRNA half life

Frozen transcript representations with task-specific linear probes, or named supervised/naive controls. Homology splits where possible; random splits for MRL-MPRA, MRL-HL-Pair and VEP. Reported mean over ten splits; Appendix C enumerates nine seeds. Table captions say ten random splits, while Methods specify homology splits where possible. This source ambiguity and the nine explicitly listed seeds are retained; exact split manifests remain unextracted. Appendix C uses micro-averaged multilabel metrics.

Aggregation: Not reported

mRNABench: A curated benchmark for mature mRNA property and function prediction · Table 6 (XML T3), data row 34, XML tr 36, column 2 (HL | Pearson R), model orthrus-large-6
Pipeline: orthrus-large-6Protocol: mRNABench variant probes MRL-HL-Pair: Paired mean ribosome load and half life
Dataset subset: mRNABench MRL-HL-Pair (mRNABench split)
0.513 ± 0.089 Pearson
unitless · higher

Uncertainty: 95% CI ± 0.089

Coverage: Not reported scored / Not reported eligible

Author-reported evaluation · Source checked
Methods, coverage and source

orthrus-large-6 on mRNABench variant probes MRL-HL-Pair: Paired mean ribosome load and half life

Frozen transcript representations with task-specific linear probes, or named supervised/naive controls. Homology splits where possible; random splits for MRL-MPRA, MRL-HL-Pair and VEP. Reported mean over ten splits; Appendix C enumerates nine seeds. Table captions say ten random splits, while Methods specify homology splits where possible. This source ambiguity and the nine explicitly listed seeds are retained; exact split manifests remain unextracted. Appendix C uses micro-averaged multilabel metrics.

Aggregation: Not reported

mRNABench: A curated benchmark for mature mRNA property and function prediction · Table 6 (XML T3), data row 34, XML tr 36, column 3 (MRL-HL-Pair | Pearson R), model orthrus-large-6
Pipeline: orthrus-large-6Protocol: mRNABench variant probes MRL-MPRA: Mean ribosome load on an MPRA library
Dataset subset: mRNABench MRL MPRA (mRNABench split)
0.629 ± 0.001 Pearson
unitless · higher

Uncertainty: 95% CI ± 0.001

Coverage: Not reported scored / Not reported eligible

Author-reported evaluation · Source checked
Methods, coverage and source

orthrus-large-6 on mRNABench variant probes MRL-MPRA: Mean ribosome load on an MPRA library

Frozen transcript representations with task-specific linear probes, or named supervised/naive controls. Homology splits where possible; random splits for MRL-MPRA, MRL-HL-Pair and VEP. Reported mean over ten splits; Appendix C enumerates nine seeds. Table captions say ten random splits, while Methods specify homology splits where possible. This source ambiguity and the nine explicitly listed seeds are retained; exact split manifests remain unextracted. Appendix C uses micro-averaged multilabel metrics.

Aggregation: Not reported

mRNABench: A curated benchmark for mature mRNA property and function prediction · Table 6 (XML T3), data row 34, XML tr 36, column 4 (MRL-MPRA | Pearson R), model orthrus-large-6
Pipeline: orthrus-large-6Protocol: mRNABench variant probes MRL: Mean ribosome load
Dataset subset: mRNABench MRL (mRNABench split)
0.427 ± 0.011 Pearson
unitless · higher

Uncertainty: 95% CI ± 0.011

Coverage: Not reported scored / Not reported eligible

Author-reported evaluation · Source checked
Methods, coverage and source

orthrus-large-6 on mRNABench variant probes MRL: Mean ribosome load

Frozen transcript representations with task-specific linear probes, or named supervised/naive controls. Homology splits where possible; random splits for MRL-MPRA, MRL-HL-Pair and VEP. Reported mean over ten splits; Appendix C enumerates nine seeds. Table captions say ten random splits, while Methods specify homology splits where possible. This source ambiguity and the nine explicitly listed seeds are retained; exact split manifests remain unextracted. Appendix C uses micro-averaged multilabel metrics.

Aggregation: Not reported

mRNABench: A curated benchmark for mature mRNA property and function prediction · Table 5 (XML T2), data row 34, XML tr 36, column 5 (MRL | Pearson R), model orthrus-large-6
Pipeline: orthrus-large-6Protocol: mRNABench variant probes Prot-Loc: Protein localisation
Dataset subset: mRNABench Prot Loc (mRNABench split)
0.393 ± 0.006 AUPRC
fraction · higher

Uncertainty: 95% CI ± 0.006

Coverage: Not reported scored / Not reported eligible

Author-reported evaluation · Source checked
Methods, coverage and source

orthrus-large-6 on mRNABench variant probes Prot-Loc: Protein localisation

Frozen transcript representations with task-specific linear probes, or named supervised/naive controls. Homology splits where possible; random splits for MRL-MPRA, MRL-HL-Pair and VEP. Reported mean over ten splits; Appendix C enumerates nine seeds. Table captions say ten random splits, while Methods specify homology splits where possible. This source ambiguity and the nine explicitly listed seeds are retained; exact split manifests remain unextracted. Appendix C uses micro-averaged multilabel metrics.

Aggregation: Not reported

mRNABench: A curated benchmark for mature mRNA property and function prediction · Table 6 (XML T3), data row 34, XML tr 36, column 6 (Prot-Loc | AUPRC), model orthrus-large-6
Pipeline: orthrus-large-6Protocol: mRNABench variant probes VEP: Variant effect prediction
Dataset subset: mRNABench VEP (mRNABench split)
0.270 ± 0.031 AUPRC
fraction · higher

Uncertainty: 95% CI ± 0.031

Coverage: Not reported scored / Not reported eligible

Author-reported evaluation · Source checked
Methods, coverage and source

orthrus-large-6 on mRNABench variant probes VEP: Variant effect prediction

Frozen transcript representations with task-specific linear probes, or named supervised/naive controls. Homology splits where possible; random splits for MRL-MPRA, MRL-HL-Pair and VEP. Reported mean over ten splits; Appendix C enumerates nine seeds. Table captions say ten random splits, while Methods specify homology splits where possible. This source ambiguity and the nine explicitly listed seeds are retained; exact split manifests remain unextracted. Appendix C uses micro-averaged multilabel metrics.

Aggregation: Not reported

mRNABench: A curated benchmark for mature mRNA property and function prediction · Table 5 (XML T2), data row 34, XML tr 36, column 6 (VEP | AUPRC), model orthrus-large-6

Source checking is not independent reproduction. Release 2026-11-21-c8bea25e5abb.

Use this pipeline

How it works, versions and access

Additional methodological details, versions and access terms have not yet been extracted into a structured profile for this pipeline. Consult the linked sources and evaluation records below.

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.

1 evidence row matching the loaded filters

Claims, original sources and review scope · Release 2026-11-21-c8bea25e5abb
Property and statementOriginal source and locationReview and provenance
Description
Linear probe, trained per mRNABench task, on frozen embeddings averaged over nucleotides from orthrus-large-6, an Orthrus checkpoint pretrained with splice and codon tracks; Orthrus is a Mamba mRNA foundation model pretrained by contrastive learning on mRNA splice isoforms and orthologs.
Individual claims
mRNABench: A curated benchmark for mature mRNA property and function prediction

Original source ↗

expansion-p3-mrnabench-2025: Appendix A, Overview of Evaluated Models (family and variant rows; naive baselines paragraph); Section 4.3, Linear Probing ('training a linear classifier on frozen embeddings'; 'averaging over per-nucleotide embeddings')

Version: preprint archived 2025-07-08
Retrieved: 2026-09-16T10:41:16.497221+00:00

source checked

["automated-source-review","ai-assisted-source-review"] · 2026-10-11

Audit details

Approved. Checked the 10 method and model descriptions against their sources (EVmutation, GEMME and RSALOR abstracts, the FramePool k-mer random forests, the RhoMax BLASSO text) and spot-checked the 49 mRNABench rows against Appendix A; all accurate. Licences on the 3 new sources match their Europe PMC records. At the reviewer's suggestion, the en dash printed in some checkpoint names is written as a hyphen in the descriptions.

Field: description

Claim: metadata-correction-91408c86b92852968f8e

Source artifact SHA-256: 79f6264ee883535203c63a313547e7c57baa85585f76b42f8d899eb17fb7e600

Hash scope: Exact retrieved primary paper artifact bytes.

Inspected artifact

Sources and history

Release 2026-11-21-c8bea25e5abb · Record review: source checked

1 source record and release historyDownload this release (gzip)
Technical metadata and record identifiers

Technical metadata for this record is tracked under stable ID mrnabench-variants-2025-method-orthrus-large-6.

Source locator: Tables 5/6, row orthrus-large-6; Appendix A (APP1), C (APP3), D (APP4)

Related records

How to cite

orthrus-large-6. Rewire biological model benchmark database, release 2026-11-21-c8bea25e5abb. https://benchmarks.rewirebio.io/database/pipeline/mrnabench-variants-2025-method-orthrus-large-6/

Release 2026-11-21-c8bea25e5abb. Rankings and values change between releases, so keep the release id in the citation. Cite the original publication as well for any value you reuse: each result links its source.

The compilation and annotations made by Rewire in this release are offered under CC BY 4.0; values and text taken from cited sources stay under the terms of those sources and are not covered by this licence. Data licence

Suggest a correction