Model type
Dilated convolutional splicing predictor
Configuration
Pangolin 1.0.2 + per-gene masking patch · mask True (P1)
Pangolin predicts splice-site strength and changes caused by genetic variants.
Sources (2)
tkzeng/Pangolin: README.md; pangolin: Journal full-text XML · Paper: Deep neural network architecture and Generating training and test sets; README.md: UsageHow it worksPangolin workflow
Conceptual summary of the documented data flow; optional inputs and configured downstream stages must be reported for a reproducible evaluation.
Sources (2)
tkzeng/Pangolin: README.md; pangolin: Journal full-text XML · Paper: Deep neural network architecture and Generating training and test sets; README.md: UsageOverview
Inputs
VCF or CSV variants, reference FASTA and matching gene annotations; custom sequence inference is also available.
Outputs
Predicted increases/decreases in splice-site strength and their positions.
Access
Official project documentation and implementation: https://github.com/tkzeng/Pangolin
Sources (2)
tkzeng/Pangolin: README.md; pangolin: Journal full-text XML · Paper: Deep neural network architecture and Generating training and test sets; README.md: Usagelimited source coverage · Automated source review, 2026-09-16. All specifications and missing details
Evaluations and results
1 evaluation · 4 results. Different protocols are not a single leaderboard.
Filter evaluations
Applied filters: All linked evaluations
Source checking is not independent reproduction. Release 2026-09-25-d40cee0abe73.
Use this model
How it works, versions and access
Related profile: Pangolin. This page retains the exact record and its evaluation context.
This configuration
8,297 of 8,324 held-out variants scored in every configuration (314 positives, 460 exon groups). The same 27 rows are excluded: 23 assembly-orientation mismatches and four outside the selected canonical transcript spans. Missing scores are not zero or negative predictions.
- record
- Pangolin 1.0.2 + per-gene masking patch · mask True (P1)
- configuration
- Not reported
- entity type
- Configuration
How it works
How it works
Pangolin predicts splice-site strength and changes caused by genetic variants. Dilated convolutional network with 16 residual blocks and skip connections; separate probability and usage outputs for heart, liver, brain and testis. The documented inputs are VCF or CSV variants, reference FASTA and matching gene annotations; custom sequence inference is also available. The output consists of predicted increases/decreases in splice-site strength and their positions.
Sources (2)
tkzeng/Pangolin: README.md; pangolin: Journal full-text XML · Paper: Deep neural network architecture and Generating training and test sets; README.md: UsageVersions and reproducibility
Pangolin implementation; gene-annotation database and selected weights must be recorded with a run. 5,000 bases upstream and downstream each output position; minimum 10,001-base input for one prediction, with 15,000-base training blocks producing 5,000 central outputs.
Sources (2)
tkzeng/Pangolin: README.md; pangolin: Journal full-text XML · Paper: Deep neural network architecture and Generating training and test sets; README.md: UsageStrengths, limitations and unresolved questions
Strengths and limitations
Strengths and considerations
- Supports custom sequences and annotation-aware variant scoring with configurable search distance.
Sources (2)
tkzeng/Pangolin: README.md; pangolin: Journal full-text XML · Paper: Deep neural network architecture and Generating training and test sets; README.md: Usage
Limitations and conditions
- Only substitutions and simple insertions/deletions are supported. The documented tool skips variants outside annotated genes, near chromosome ends, inconsistent with the reference or beyond supported deletion lengths.
Sources (2)
tkzeng/Pangolin: README.md; pangolin: Journal full-text XML · Paper: Deep neural network architecture and Generating training and test sets; README.md: Usage
Profile review details
Inspected pinned official documentation, relevant implementation files and named primary-paper sections. Claims are limited to those artifacts. Remaining field extraction and identity conflicts are explicit; no new performance claims, model runs or human review are implied.
Stable record: discovery-model-pangolinSpecifications
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 |
|---|---|
| Model type | Dilated convolutional splicing predictorSources (2)tkzeng/Pangolin: README.md; pangolin: Journal full-text XML · Paper: Deep neural network architecture and Generating training and test sets; README.md: Usage |
| Architecture | Dilated convolutional network with 16 residual blocks and skip connections; separate probability and usage outputs for heart, liver, brain and testis.Sources (2)tkzeng/Pangolin: README.md; pangolin: Journal full-text XML · Paper: Deep neural network architecture and Generating training and test sets; README.md: Usage |
| Inputs | VCF or CSV variants, reference FASTA and matching gene annotations; custom sequence inference is also available.Sources (2)tkzeng/Pangolin: README.md; pangolin: Journal full-text XML · Paper: Deep neural network architecture and Generating training and test sets; README.md: Usage |
| Outputs | Predicted increases/decreases in splice-site strength and their positions.Sources (2)tkzeng/Pangolin: README.md; pangolin: Journal full-text XML · Paper: Deep neural network architecture and Generating training and test sets; README.md: Usage |
| Parameters | The reviewed architecture section specifies the dilated residual network, but does not give a complete parameter total for the released ensemble. · Not reported in inspected sourcesSources (2)tkzeng/Pangolin: README.md; pangolin: Journal full-text XML · Paper: Deep neural network architecture and Generating training and test sets; README.md: Usage |
| Known versions | Pangolin implementation; gene-annotation database and selected weights must be recorded with a run.Sources (2)tkzeng/Pangolin: README.md; pangolin: Journal full-text XML · Paper: Deep neural network architecture and Generating training and test sets; README.md: Usage |
| Training data | Human, rhesus macaque, mouse and rat sequence/splicing data. Human test chromosomes 1, 3, 5, 7 and 9 are held out, with homologous training genes filtered using Ensembl BioMart.Sources (2)tkzeng/Pangolin: README.md; pangolin: Journal full-text XML · Paper: Deep neural network architecture and Generating training and test sets; README.md: Usage |
| Training cutoff | Training annotations are GENCODE 34 (human), Ensembl 100 (rhesus), GENCODE M25 (mouse) and Ensembl 101 (rat). These component releases do not establish one latest RNA-seq collection date.Sources (2)tkzeng/Pangolin: README.md; pangolin: Journal full-text XML · Paper: Deep neural network architecture and Generating training and test sets; README.md: Usage |
| Context limits | 5,000 bases upstream and downstream each output position; minimum 10,001-base input for one prediction, with 15,000-base training blocks producing 5,000 central outputs.Sources (2)tkzeng/Pangolin: README.md; pangolin: Journal full-text XML · Paper: Deep neural network architecture and Generating training and test sets; README.md: Usage |
| Weights licence | Separate checkpoint-distribution terms are not stated in the inspected release documentation and licence material. The source-code licence alone is not recorded as an explicit weight grant. · Not reported in inspected sourcesSources (3)tkzeng/Pangolin: README.md; pangolin: Journal full-text XML; tkzeng/Pangolin: LICENSE · Paper: Deep neural network architecture and Generating training and test sets; README.md: Usage; LICENSE: licence text |
| Access | Official project documentation and implementation: https://github.com/tkzeng/PangolinSources (2)tkzeng/Pangolin: README.md; pangolin: Journal full-text XML · Paper: Deep neural network architecture and Generating training and test sets; README.md: Usage |
| Code licence | GPL-3.0; inspect the pinned licence and any file-specific terms.Sourcestkzeng/Pangolin: LICENSE · LICENSE: licence text |
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.
5 evidence rows matching the loaded filters
| Property and statement | Original source and location | Review and provenance |
|---|---|---|
| Relationship: variant of discovery-model-pangolin Individual claims | MFASS matched canonical annotation v1: exclusion-verification.json manifest-v1.json: study and conditions Shared locator for this statement’s cited sources; not a separate locator for each citation. Version: 4be7a98e2553fa2378c29625b13eb3e8ac2e58fb | source checked automated execution evidence review · 2026-09-25T11:24:56Z Audit detailsComputational source, coverage and metric checks only; not independent human review or published-score reproduction. Field: Claim: rewire-mfass-matched-v1-configuration-p1-variant-of-discovery-model-pangolin Source artifact SHA-256: Hash scope: Hash scope not separately documented; inspect source record |
| Relationship: variant of discovery-model-pangolin Individual claims | MFASS matched canonical annotation v1: manifest-v1.json manifest-v1.json: study and conditions Shared locator for this statement’s cited sources; not a separate locator for each citation. Version: 093fd1ae198c80ce34408d84d6543bca4fc538f2 | source checked automated execution evidence review · 2026-09-25T11:24:56Z Audit detailsComputational source, coverage and metric checks only; not independent human review or published-score reproduction. Field: Claim: rewire-mfass-matched-v1-configuration-p1-variant-of-discovery-model-pangolin Source artifact SHA-256: Hash scope: Hash scope not separately documented; inspect source record |
| Relationship: variant of discovery-model-pangolin Individual claims | MFASS matched canonical annotation v1: provenance.json manifest-v1.json: study and conditions Shared locator for this statement’s cited sources; not a separate locator for each citation. Version: 093fd1ae198c80ce34408d84d6543bca4fc538f2 | source checked automated execution evidence review · 2026-09-25T11:24:56Z Audit detailsComputational source, coverage and metric checks only; not independent human review or published-score reproduction. Field: Claim: rewire-mfass-matched-v1-configuration-p1-variant-of-discovery-model-pangolin Source artifact SHA-256: Hash scope: Hash scope not separately documented; inspect source record |
| Relationship: variant of discovery-model-pangolin Individual claims | MFASS matched canonical annotation v1: report.json manifest-v1.json: study and conditions Shared locator for this statement’s cited sources; not a separate locator for each citation. Version: 093fd1ae198c80ce34408d84d6543bca4fc538f2 | source checked automated execution evidence review · 2026-09-25T11:24:56Z Audit detailsComputational source, coverage and metric checks only; not independent human review or published-score reproduction. Field: Claim: rewire-mfass-matched-v1-configuration-p1-variant-of-discovery-model-pangolin Source artifact SHA-256: Hash scope: Hash scope not separately documented; inspect source record |
| Relationship: variant of discovery-model-pangolin Individual claims | MFASS matched canonical annotation v1: verification.json manifest-v1.json: study and conditions Shared locator for this statement’s cited sources; not a separate locator for each citation. Version: 093fd1ae198c80ce34408d84d6543bca4fc538f2 | source checked automated execution evidence review · 2026-09-25T11:24:56Z Audit detailsComputational source, coverage and metric checks only; not independent human review or published-score reproduction. Field: Claim: rewire-mfass-matched-v1-configuration-p1-variant-of-discovery-model-pangolin 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 2026-09-25-d40cee0abe73 · Record review: source checked
5 source records and release history
- MFASS matched canonical annotation v1: report.json · Original source · 093fd1ae198c80ce34408d84d6543bca4fc538f2
- MFASS matched canonical annotation v1: manifest-v1.json · Original source · 093fd1ae198c80ce34408d84d6543bca4fc538f2
- MFASS matched canonical annotation v1: verification.json · Original source · 093fd1ae198c80ce34408d84d6543bca4fc538f2
- MFASS matched canonical annotation v1: provenance.json · Original source · 093fd1ae198c80ce34408d84d6543bca4fc538f2
- MFASS matched canonical annotation v1: exclusion-verification.json · Original source · 4be7a98e2553fa2378c29625b13eb3e8ac2e58fb
Technical metadata and extraction receipts
Stable ID: rewire-mfass-matched-v1-configuration-p1
- areas
- dna-genomes
- condition
- P1
- configuration
- tool: pangolin; mask: True; stem: pangolin-gencode44-canonical-maskTrue
- distance
- 50
- annotation
- GENCODE v44 canonical transcripts
- code
- pangolin source sha256: 239c6ba9b1883d712d4856985c6eaa07e5b67f99790a718f93d6316ad076c083; pangolin source identity: pangolin-5cf94b8-mask-per-gene-1; pangolin version: 1.0.2+rewire.maskpergene1; pangolin patch: id: pangolin-5cf94b8-mask-per-gene-1; patch sha256: d1c9fb69245f9d256e2547fd8bf61547686cfc3caf6c40c1351ab2d6997cb454; upstream revision: 5cf94b8db938c658391b4305cd7ce33297d44ff7; upstream source sha256: f3f25c4febf64d01ef42f967dc5cf10f6856bf3aa92d26f09803ad408169da9b; pangolin files: init .py: sha256: e3b0c44298fc1c149afbf4c8996fb92427ae41e4649b934ca495991b7852b855; git blob: e69de29bb2d1d6434b8b29ae775ad8c2e48c5391; model.py: sha256: 4a1c5c2570aafe1452bb43332255321677e6c6c817adf84b9dd438e3ca4be6f8; git blob: 11dfb436b36f9b5783874b2db636a1788bfab619; models/final.1.0.3.v2: sha256: f0478fab173b75f7f7e9fe96688bad6c50fa4a46d70557f423b110caaf565501; git blob: fbe03218df2f781a9713b05bfa42001ae548782a; models/final.2.0.3.v2: sha256: c4c6bb4880fa6fb28b14182ae3ea0600edb07056158f55325b5e6e6e48fc9f26; git blob: ec4ee767bc83fa0397222f42c0da210a47f13fbb; models/final.3.0.3.v2: sha256: ec685a6e7105a4486c1f89a005458a13deb3fe7171f13d434f4877e386d10676; git blob: f31cbae821e38bc4361216c8e87c8db13fd5a9c0; models/final.1.2.3.v2: sha256: 559c05de3e1ce65c2515ca3e92ef85edb0ec2e47686ca58060e25891ce06eb3a; git blob: 39e8756faf1e427cd9a8c6c691532b3683d8916a; models/final.2.2.3.v2: sha256: 48758ba8b95eee9aa9feea52672ef06ca1b34111299c27f8a710f734d8b9aae5; git blob: 6fd0a31789c7de788227e103f33d697ad538def6; models/final.3.2.3.v2: sha256: 7cb576c2b24db4fdd6970c4ca4fb7c20ae1b1d8ae80645ebbe689848b5743129; git blob: c7f2ae10caab8f5bae167309b2b7d3e6905871e5; models/final.1.4.3.v2: sha256: c50b12e0c0af776d5674ca5e346493f8265783494d4df383364de9c1136657f6; git blob: 0af4fadb3f3dd54c49d41042befefa19fd4bd32a; models/final.2.4.3.v2: sha256: e03303bed4fd6f135ec0f6c1b192cce954ea42d0646f44d17b4a6fbb2b1f610e; git blob: 8c003df629c9f1e8072000f8bbc3eeab94850b6c; models/final.3.4.3.v2: sha256: 9476d2e25520d7ff15bece0cd5d3b657e3b1dd3cc5fcab1d9c3b62bea7a0c5b6; git blob: 9ed4fad50ead41761073cdff3d9acafc6f78bdd4; models/final.1.6.3.v2: sha256: 2aae563fa18a8a9b6699c6c96e0d32b8ec7543f8f805fb3bc9de77302cc9f66e; git blob: 98e5539fe339385e9c8b19d2786c4f8c3f186a3b; models/final.2.6.3.v2: sha256: 7d3c0b1b2a60067b940dec315567874fbc8bcd322f1b7c76bf969f51f0f53f7f; git blob: 6a650d629fae6aee8171708241a0f580bb368cfd; models/final.3.6.3.v2: sha256: 756e7721a382cace24e9bfea5b543af5623f2487d9a3efe7385e9c76367005fd; git blob: 4a8e9e33d5832770e6563d0229bdb4bcae5b8398; pangolin upstream files verified: true; pangolin upstream file mismatches: None recorded
- resources sha256
- cohort: 389702ff4c647d7ce10a90092a6fa811ae777d15997baf39ce9aae0346247bd0; split: 999ebcb7e63a5c5eaa8780fa468e59ac1f934260ad50102814174c396317f052; fasta: e49b92b3e4f321bf254c042f25b726d9931c4d74c7523e8b6bb530e63b0cfd4b; fai: a2c323ea4cff34d7123ace4578f7e122b2d2f5a22f40dc23eb8b97d17723d169; spliceai annotation: 2383cd148b292b9a6af56da160c9c142ddeb5720b1e9ef7c5966a1744f684c4d; pangolin db: 84ae1a18d9d7a32e14e3fe9c8705256a28cd9bde0a55bc46200cbe77178deb27
- source locator
- manifest-v1.json: conditions.P1; code.pangolin
- limitations
- 8,297 of 8,324 held-out variants scored in every configuration (314 positives, 460 exon groups). The same 27 rows are excluded: 23 assembly-orientation mismatches and four outside the selected canonical transcript spans. Missing scores are not zero or negative predictions.; Exploratory comparison: prior results were known. Paired contrast intervals are unadjusted and do not establish a universal model ranking.; Pangolin uses the recorded per-gene masking patch; these are exact configurations, not unqualified upstream model scores.; Precision at 100 is sensitive to tied-score ordering, especially P1. Numerical source checking is automated, not human review or independent reproduction.; Assembly-orientation issue reported at https://github.com/KosuriLab/MFASS/issues/1. Original v1 outputs remain unchanged; corrections require a new version.