Model type
Graph-aware transformer pipeline; this record is the paper-specific evaluated configuration.
Pipeline
GSMFormer-PPI + ProstT5
GSMFormer-PPI combines molecular surfaces, structural graphs and ProstT5 sequence features for interaction prediction.
Sources
Multimodal graph, surface, and language-based model for protein protein interaction prediction · Results/Ablation experiments for input modalities (paragraph 1); Conclusions (paragraph 1)How it worksEvaluated procedure (conceptual)
Conceptual input–method–output guide. Check the procedure text and linked evaluation for fitted components, additional inputs and exact settings.
Sources
Multimodal graph, surface, and language-based model for protein protein interaction prediction · Methods/Model architecture/Integration of protein structural and sequence features (paragraph 2); Methods/Model architecture/Linear projectors (paragraph 1)At a glance
Biological inputs
Protein surface and 3D structure features plus ProstT5 sequence embeddings
Sources
Multimodal graph, surface, and language-based model for protein protein interaction prediction · Methods/Graph representation of 3D protein structure (paragraph 2); Discussion (paragraph 7)Outputs
Protein–protein interaction predictions
Sources
Multimodal graph, surface, and language-based model for protein protein interaction prediction · Methods/Graph representation of 3D protein structure/Graph Convolutional Neural Network (GCN) (paragraph 1); Methods/Model architecture/Integration of protein structural and sequence features (paragraph 2)limited source coverage · Automated source review, 2026-09-16. All specifications and missing details
Evaluations and results
Release 2026-09-17-d277315f7d76 · 1 evaluation · 1 metric row. Different protocols are not a single leaderboard.
| Metric and finding | Coverage and uncertainty | Evidence |
|---|---|---|
| GSMFormer-PPI + ProstT5: protein-protein interaction prediction Pipeline: GSMFormer-PPI + ProstT5Task: protein-protein interaction predictionDataset: paper PPI test set ProstT5 embeddings as graph node features Author-reported evaluation · Evaluation metadata: needs review | ||
| 0.988 AUROC Unit: fraction · Direction: unknown | Uncertainty: not reported in legacy extract Scored: Not reported · Eligible: Not reported | source checkedMultimodal graph, surface, and language-based model for protein protein interaction prediction · Table 6, ProstT5 embedding row, AUROC column Source checking is not independent reproduction. |
How it works
How the evaluated method works
MaSIF extracts physicochemical surface descriptors, graph convolutions process protein structure, and a transformer with linear projectors fuses these with residue-level embeddings.
Sources
Multimodal graph, surface, and language-based model for protein protein interaction prediction · Methods/Model architecture/Integration of protein structural and sequence features (paragraph 2); Methods/Model architecture/Linear projectors (paragraph 1)What was evaluated
The linked evaluation record identifies GSMFormer-PPI + ProstT5: protein-protein interaction prediction. Its dataset, split, adaptation and evidence origin remain attached to the reported results.
Sources
Multimodal graph, surface, and language-based model for protein protein interaction prediction · The named evaluation’s methods and comparison table; exact preserved evaluation IDs: evaluation-b2-gsmformer-ppi-2026Strengths and limitations
Strengths and considerations
- Ablation studies examine the contribution of surface features and multimodal fusion.
Sources
Multimodal graph, surface, and language-based model for protein protein interaction prediction · Abstract (paragraph 1); Conclusions (paragraph 1)
Limitations and conditions
- Cross-dataset comparisons of separately reported scores are not a matched head-to-head benchmark; structural inputs confer information beyond sequence-only models.
Sources
Multimodal graph, surface, and language-based model for protein protein interaction prediction · Abstract (paragraph 1); Results/Performance of GSMFormer-PPI on Baranwal’s dataset (paragraph 1)
Profile review details
Primary full text and the available official implementation/model documentation were inspected. Explanatory claims are source-backed; unresolved exact-configuration metadata is labelled explicitly. This is automated review, not a human review or independent benchmark reproduction.
Stable record: reported-model-73ae07fb5be204Specifications
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 | Graph-aware transformer pipeline; this record is the paper-specific evaluated configuration.SourcesMultimodal graph, surface, and language-based model for protein protein interaction prediction · Methods/Model architecture/Integration of protein structural and sequence features (paragraph 2); Methods/Model architecture/Linear projectors (paragraph 1) |
| Architecture / procedure | MaSIF extracts physicochemical surface descriptors, graph convolutions process protein structure, and a transformer with linear projectors fuses these with residue-level embeddings.SourcesMultimodal graph, surface, and language-based model for protein protein interaction prediction · Methods/Model architecture/Integration of protein structural and sequence features (paragraph 2); Methods/Model architecture/Linear projectors (paragraph 1) |
| Biological inputs | Protein surface and 3D structure features plus ProstT5 sequence embeddingsSourcesMultimodal graph, surface, and language-based model for protein protein interaction prediction · Methods/Graph representation of 3D protein structure (paragraph 2); Discussion (paragraph 7) |
| Outputs | Protein–protein interaction predictionsSourcesMultimodal graph, surface, and language-based model for protein protein interaction prediction · Methods/Graph representation of 3D protein structure/Graph Convolutional Neural Network (GCN) (paragraph 1); Methods/Model architecture/Integration of protein structural and sequence features (paragraph 2) |
| Parameters | Table 3 reports 2.8M parameters for GSMFormer-PPI; the pretrained ProstT5 feature extractor is a separate component.SourcesMultimodal graph, surface, and language-based model for protein protein interaction prediction · Table3; GSMFormer-PPI row, Parameters column |
| Known versions / configuration | GSMFormer-PPI + ProstT5 is the comparison-table label; that label does not specify an immutable weight revision. · Not reported in inspected sourcesSourcesMultimodal graph, surface, and language-based model for protein protein interaction prediction · Model identification in the comparison table and corresponding Methods; immutable checkpoint revision is not supplied by the table label. |
| Training data / fitting | PINDER ML-ready splits with interface- and sequence-similarity controls, including the paper’s stated 30% sequence-identity threshold.SourcesMultimodal graph, surface, and language-based model for protein protein interaction prediction · Methods/Dataset/Splits (paragraph 1); Methods/Dataset/Splits (paragraph 2) |
| Context limits | Curated dimer chains contain 35–300 residues; embeddings are padded to the longest sequence in each batch.SourcesMultimodal graph, surface, and language-based model for protein protein interaction prediction · Methods/Protein sequence feature extraction (paragraph 3); Methods/Graph representation of 3D protein structure/Graph Convolutional Neural Network (GCN) (paragraph 2) |
| Access | Official study implementation and usage documentation: https://github.com/ChervovNikita/gsmformer-ppi/blob/db9886e8b295f35b544a5703659e2a9115ce9e22/README.md. This pinned documentation revision is not automatically the evaluated weight revision.SourcesChervovNikita/gsmformer-ppi README.md · README.md; installation, model download and usage instructions |
| Code licence | No explicit code licence was established from the paper’s availability statement and inspected repository-root documentation. · Not reported in inspected sourcesSourcesChervovNikita/gsmformer-ppi README.md · README.md and repository-root licence-file search |
| Weights licence | The inspected model-access documentation does not explicitly identify terms for this exact evaluated checkpoint or fitted head; repository code terms are shown separately. · Not reported in inspected sourcesSourcesChervovNikita/gsmformer-ppi README.md · README.md; checkpoint/access documentation and licence scope |
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.
19 evidence rows matching the loaded filters
| Property and statement | Original source and location | Review and provenance |
|---|---|---|
| Diagram caption Conceptual input–method–output guide. Check the procedure text and linked evaluation for fitted components, additional inputs and exact settings. Individual claims | Multimodal graph, surface, and language-based model for protein protein interaction prediction Methods/Model architecture/Integration of protein structural and sequence features (paragraph 2); Methods/Model architecture/Linear projectors (paragraph 1) Version: journal full text in PMC | source checked automated source review · 2026-09-16 Audit detailsPrimary full text and the available official implementation/model documentation were inspected. Explanatory claims are source-backed; unresolved exact-configuration metadata is labelled explicitly. This is automated review, not a human review or independent benchmark reproduction. Field: Source artifact SHA-256: Hash scope: Hash scope not separately documented; inspect source record |
| Diagram steps ["Protein surface and 3D structure features plus ProstT5 sequence embeddings","GSMFormer-PPI + ProstT5","Protein–protein interaction predictions"] Individual claims | Multimodal graph, surface, and language-based model for protein protein interaction prediction Methods/Model architecture/Integration of protein structural and sequence features (paragraph 2); Methods/Model architecture/Linear projectors (paragraph 1) Version: journal full text in PMC | source checked automated source review · 2026-09-16 Audit detailsPrimary full text and the available official implementation/model documentation were inspected. Explanatory claims are source-backed; unresolved exact-configuration metadata is labelled explicitly. This is automated review, not a human review or independent benchmark reproduction. Field: Source artifact SHA-256: Hash scope: Hash scope not separately documented; inspect source record |
| Diagram title Evaluated procedure (conceptual) Individual claims | Multimodal graph, surface, and language-based model for protein protein interaction prediction Methods/Model architecture/Integration of protein structural and sequence features (paragraph 2); Methods/Model architecture/Linear projectors (paragraph 1) Version: journal full text in PMC | source checked automated source review · 2026-09-16 Audit detailsPrimary full text and the available official implementation/model documentation were inspected. Explanatory claims are source-backed; unresolved exact-configuration metadata is labelled explicitly. This is automated review, not a human review or independent benchmark reproduction. Field: Source artifact SHA-256: Hash scope: Hash scope not separately documented; inspect source record |
| Model type Graph-aware transformer pipeline; this record is the paper-specific evaluated configuration. Individual claims | Multimodal graph, surface, and language-based model for protein protein interaction prediction Methods/Model architecture/Integration of protein structural and sequence features (paragraph 2); Methods/Model architecture/Linear projectors (paragraph 1) Version: journal full text in PMC | source checked automated source review · 2026-09-16 Audit detailsPrimary full text and the available official implementation/model documentation were inspected. Explanatory claims are source-backed; unresolved exact-configuration metadata is labelled explicitly. This is automated review, not a human review or independent benchmark reproduction. Field: Source artifact SHA-256: Hash scope: Hash scope not separately documented; inspect source record |
| Architecture / procedure MaSIF extracts physicochemical surface descriptors, graph convolutions process protein structure, and a transformer with linear projectors fuses these with residue-level embeddings. Individual claims | Multimodal graph, surface, and language-based model for protein protein interaction prediction Methods/Model architecture/Integration of protein structural and sequence features (paragraph 2); Methods/Model architecture/Linear projectors (paragraph 1) Version: journal full text in PMC | source checked automated source review · 2026-09-16 Audit detailsPrimary full text and the available official implementation/model documentation were inspected. Explanatory claims are source-backed; unresolved exact-configuration metadata is labelled explicitly. This is automated review, not a human review or independent benchmark reproduction. Field: Source artifact SHA-256: Hash scope: Hash scope not separately documented; inspect source record |
| Weights licence The inspected model-access documentation does not explicitly identify terms for this exact evaluated checkpoint or fitted head; repository code terms are shown separately. Individual claims | ChervovNikita/gsmformer-ppi README.md README.md; checkpoint/access documentation and licence scope Version: db9886e8b295f35b544a5703659e2a9115ce9e22 | unreported automated source review · 2026-09-16 Audit detailsPrimary full text and the available official implementation/model documentation were inspected. Explanatory claims are source-backed; unresolved exact-configuration metadata is labelled explicitly. This is automated review, not a human review or independent benchmark reproduction. Field: Source artifact SHA-256: Hash scope: Hash scope not separately documented; inspect source record |
| Biological inputs Protein surface and 3D structure features plus ProstT5 sequence embeddings Individual claims | Multimodal graph, surface, and language-based model for protein protein interaction prediction Methods/Graph representation of 3D protein structure (paragraph 2); Discussion (paragraph 7) Version: journal full text in PMC | source checked automated source review · 2026-09-16 Audit detailsPrimary full text and the available official implementation/model documentation were inspected. Explanatory claims are source-backed; unresolved exact-configuration metadata is labelled explicitly. This is automated review, not a human review or independent benchmark reproduction. Field: Source artifact SHA-256: Hash scope: Hash scope not separately documented; inspect source record |
| Outputs Protein–protein interaction predictions Individual claims | Multimodal graph, surface, and language-based model for protein protein interaction prediction Methods/Graph representation of 3D protein structure/Graph Convolutional Neural Network (GCN) (paragraph 1); Methods/Model architecture/Integration of protein structural and sequence features (paragraph 2) Version: journal full text in PMC | source checked automated source review · 2026-09-16 Audit detailsPrimary full text and the available official implementation/model documentation were inspected. Explanatory claims are source-backed; unresolved exact-configuration metadata is labelled explicitly. This is automated review, not a human review or independent benchmark reproduction. Field: Source artifact SHA-256: Hash scope: Hash scope not separately documented; inspect source record |
| Parameters Table 3 reports 2.8M parameters for GSMFormer-PPI; the pretrained ProstT5 feature extractor is a separate component. Individual claims | Multimodal graph, surface, and language-based model for protein protein interaction prediction Table3; GSMFormer-PPI row, Parameters column Version: journal full text in PMC | source checked automated source review · 2026-09-16 Audit detailsPrimary full text and the available official implementation/model documentation were inspected. Explanatory claims are source-backed; unresolved exact-configuration metadata is labelled explicitly. This is automated review, not a human review or independent benchmark reproduction. Field: Source artifact SHA-256: Hash scope: Hash scope not separately documented; inspect source record |
| Known versions / configuration GSMFormer-PPI + ProstT5 is the comparison-table label; that label does not specify an immutable weight revision. Individual claims | Multimodal graph, surface, and language-based model for protein protein interaction prediction Model identification in the comparison table and corresponding Methods; immutable checkpoint revision is not supplied by the table label. Version: journal full text in PMC | unreported automated source review · 2026-09-16 Audit detailsPrimary full text and the available official implementation/model documentation were inspected. Explanatory claims are source-backed; unresolved exact-configuration metadata is labelled explicitly. This is automated review, not a human review or independent benchmark reproduction. Field: Source artifact SHA-256: Hash scope: Hash scope not separately documented; inspect source record |
Sources and history
Release 2026-09-17-d277315f7d76 · Record review: needs review
2 source records and release history
- Multimodal graph, surface, and language-based model for protein protein interaction prediction · Original source · journal full text in PMC
- ChervovNikita/gsmformer-ppi README.md · Original source · db9886e8b295f35b544a5703659e2a9115ce9e22
Technical metadata and extraction receipts
Stable ID: reported-model-73ae07fb5be204
- areas
- proteins-complexes
- entity level
- method
- version
- not stated in table
- reported name
- GSMFormer-PPI + ProstT5
- historical missing metadata
- checkpoint revision: not_reported_in_legacy_extract; training data: not_reported_in_legacy_extract; licence: 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
- model
- entity classification
- review date: 2026-09-17; rationale: This record identifies a composed analysis workflow with separately identifiable upstream models, representations or tools and a downstream prediction/scoring procedure. Results belong to that complete composition rather than to an upstream model alone.; source ids: gsmformer-ppi-2026; source locator: Methods/Model architecture/Integration of protein structural and sequence features (paragraph 2); Methods/Model architecture/Linear projectors (paragraph 1) | Results/Ablation experiments for input modalities (paragraph 1); Conclusions (paragraph 1); ambiguities: This is the paper-specific pipeline identity; unspecified component checkpoints or implementation versions are not inferred from its name.