GTEx tissue-specific splicing prediction, Splice-15class (You et al. 2026 Supplementary Table 5)
Splice site usage across 15 human tissues: every method scores all splice sites within the central 500 bp of the input sequence.
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.
7 recorded evaluations, 7 metric rows. A comparison chart has not yet been validated for these results. The table retains the individual findings and their sources.
Results
Results are available, but no reviewed comparison panel is linked in this release.
All evaluations
7 evaluations · 7 results. Different protocols are not a single leaderboard.
Protocol: GTEx tissue-specific splicing prediction, Splice-15class (You et al. 2026 Supplementary Table 5) · Dataset: GTEx-derived tissue-specific splice site usage dataset (You et al. 2026)
Sorted by AUPRC (GTEx splicing task, Splice-15class, averaged over categories) (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 7 result rows with coverage, uncertainty and sources
Filter evaluations
Applied filters: All linked evaluations
Source checking is not independent reproduction. Release 2026-10-22-1b7cf027e4f6.
Methods and evaluation design
Procedure, tasks and evaluated configurations
Recorded evaluations
Each evaluation records what was tested and under which conditions.
- NucleotideTransformer on the GTEx splicing task, Splice-15class
- RNAErnie on the GTEx splicing task, Splice-15class
- RNAFM on the GTEx splicing task, Splice-15class
- RNAMSM on the GTEx splicing task, Splice-15class
- SpliceBERT on the GTEx splicing task, Splice-15class
- SpTransformer on the GTEx splicing task, Splice-15class
- UTRLM on the GTEx splicing task, Splice-15class
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
- 1
- External evaluations
- 6
Literature evidence is not a Rewire measurement. Executed but unpublished runs and private review status are not included.
Null control
Proposed control: requires review
Training-set class prior where supervised fitting is permitted
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
Regularised classifier on simple permitted features, or protocol's conventional reference
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 2026-10-22-1b7cf027e4f6. 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.
Strengths, limitations and unresolved questions
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
| Property and statement | Original source and location | Review and provenance |
|---|
No evidence rows match these filters. Choose another scope or clear the search.
Sources and history
Release 2026-10-22-1b7cf027e4f6 · Record review: source checked
2 source records and release history
- Benchmarking pre-trained genomic language models for RNA sequence-related predictive applications · Original source · Nat Commun 17:223, published online 2025-12-07; PMC12780105 full-text XML
- You et al. 2026, Supplementary Information · Original source · 41467_2025_66899_MOESM1_ESM.pdf inside the Europe PMC supplementaryFiles zip (outer zip 9a0d0a8d8643872ab265e3e6ed7eea944563da4861bf764f97c68fc6aa502525, assembled per request; the member hash is the bound artifact)
Technical metadata and extraction receipts
Stable ID: gfm2-20261011-you2026-protocol-gtex-splice-15class
- areas
- dna-genomes
- contexts
- research
- protocol
- Score all potential splice sites within the centre 500 bp of each input sequence from the GTEx-derived dataset; fine-tune and test three times with different random initialisations and report mean and standard deviation of average PR-AUC and average top-k accuracy.
- version
- Supplementary Table 5, block 'Splice-15class'
- metric direction
- higher
- unit
- fraction
- aggregation
- Mean over three fine-tuning runs; metrics averaged over the task's output categories
- replicates
- Three random initialisations
- metric implementation
- Average PR-AUC and SpliceAI-style top-k accuracy, chosen because splice sites are sparse and the classes extremely imbalanced (Results)
- limitations
- Fine-tuned and tested three times with different random initialisations; the printed spread is the standard deviation over those three runs (Methods, 'To ensure the reliability of our findings').; All methods are given a 500 bp input so that models with different context windows are comparable; SpTransformer and SpliceAI were designed for 1,000 to 5,000 nucleotide inputs (Results).; Each language model is fine-tuned with an appended convolution head, so the comparison measures the adapted model, not zero-shot likelihood.; Values are read from the Supplementary Information tables behind the figures; the main-text figures are not read.; Not every method appears in every block: the 15- and 53-class blocks list seven methods and the pooled acceptor and donor block eight.; Each model's reported values come from the epoch with the highest average PR-AUC across classes (Methods); the dataset is described only as a train and test split, and no validation split is named.
- source locator
- Supplementary Table 5, block 'Splice-15class'; Methods
Related records
- uses data: GTEx-derived tissue-specific splice site usage dataset (You et al. 2026)
- assessment: NucleotideTransformer on the GTEx splicing task, Splice-15class
- assessment: RNAErnie on the GTEx splicing task, Splice-15class
- assessment: RNAFM on the GTEx splicing task, Splice-15class
- assessment: RNAMSM on the GTEx splicing task, Splice-15class
- assessment: SpliceBERT on the GTEx splicing task, Splice-15class
- assessment: SpTransformer on the GTEx splicing task, Splice-15class
- assessment: UTRLM on the GTEx splicing task, Splice-15class