Model

GEARS

GEARS predicts transcriptional responses to single- and multi-gene perturbations from single-cell perturbation screens.

Sources (3)

snap-stanford/GEARS: README.md; snap-stanford/GEARS: gears/model.py; gears: Journal full-text XML · GEARS paper Methods: Overview, Gene coexpression graph encoder, GO graph, Cross-gene effects and gene-specific decoder; gears/model.py: GEARS_Model

1 evaluation · 2 metric rows

How it worksGEARS workflow

Conceptual summary of the documented data flow; optional inputs and configured downstream stages must be reported for a reproducible evaluation.

Sources (3)

Overview

Model type

Graph-based perturbation-response predictor

Sources (3)

Inputs

Single-cell expression data, perturbation labels and the graph resources used by the configured model.

Sources (3)

Outputs

Predicted gene-expression responses and genetic-interaction analyses.

Sources (3)

Access

Official project documentation and implementation: https://github.com/snap-stanford/GEARS

Sources (3)

limited source coverage · Automated source review, 2026-09-16. All specifications and missing details

Evaluations and results

1 evaluation · 2 metric rows. Different protocols are not a single leaderboard.

Filter evaluations

Tested model or methodProtocolDatasetEvaluation setupMetricEvidence origin

Applied filters: All linked evaluations

Exact evaluated configurations and original reported results
Tested configuration	Protocol and dataset	Finding	Evidence and details
Configuration: GEARS	Protocol: GEARS Norman2019 CPA-control comparison MSE: Norman2019 perturbation-response MSE Dataset subset: Norman et al.2019 perturbation data used in GEARS Supplementary Table6 (GEARS Norman2019 CPA-control comparison split)	0.216 ± 0.053 mse squared_expression_units_unreported · lower Uncertainty: type: unreported; reported spread: 0.053; note: Printed ± spread; SD/SE/CI type not established by inspected table caption. Coverage: Not reported scored / Not reported eligible	Author-reported evaluation · source checked Methods, coverage and source GEARS on GEARS Norman2019 CPA-control comparison MSE: Norman2019 perturbation-response MSE Predict post-perturbation expression on Norman2019. MSE compares predicted and true expression; Pearson DE compares predicted versus true change over unperturbed controls (Supplementary Table1). Exact Table6 scoring gene subset and split manifest remain unextracted. Aggregation: Not reported GEARS primary supplement, Table6 · Supplementary Table6, printed p34 / PDF p35, row GEARS, column MSE
Configuration: GEARS	Protocol: GEARS Norman2019 CPA-control comparison Pearson DE: Norman2019 perturbation-response Pearson DE Dataset subset: Norman et al.2019 perturbation data used in GEARS Supplementary Table6 (GEARS Norman2019 CPA-control comparison split)	0.556 ± 0.030 pearson_delta_expression correlation · higher Uncertainty: type: unreported; reported spread: 0.030; note: Printed ± spread; SD/SE/CI type not established by inspected table caption. Coverage: Not reported scored / Not reported eligible	Author-reported evaluation · source checked Methods, coverage and source GEARS on GEARS Norman2019 CPA-control comparison Pearson DE: Norman2019 perturbation-response Pearson DE Predict post-perturbation expression on Norman2019. MSE compares predicted and true expression; Pearson DE compares predicted versus true change over unperturbed controls (Supplementary Table1). Exact Table6 scoring gene subset and split manifest remain unextracted. Aggregation: Not reported GEARS primary supplement, Table6 · Supplementary Table6, printed p34 / PDF p35, row GEARS, column Pearson DE

Source checking is not independent reproduction. Release 2026-09-23-2b89723c6dd9.

Use this model

How it works, versions and access

Related profile: GEARS. This page retains the exact record and its evaluation context.

Versions and evaluated configurations

GEARS · alias of
GEARS · family

How it works

GEARS predicts transcriptional responses to single- and multi-gene perturbations from single-cell perturbation screens. Two graph encoders represent gene coexpression and Gene Ontology perturbation similarity. Perturbation embeddings are composed with gene embeddings, then a cross-gene network and gene-specific decoders predict expression changes. The documented inputs are single-cell expression data, perturbation labels and the graph resources used by the configured model. The output consists of predicted gene-expression responses and genetic-interaction analyses.

Sources (3)

Versions and reproducibility

The README describes v0.1.1 updates; a specific trained checkpoint must be recorded separately. A gene-expression vector and perturbation set over the configured gene inventory; no fixed nucleotide or amino-acid token window.

Sources (3)

Strengths, limitations and unresolved questions

Strengths and limitations

Strengths and considerations

Provides documented dataset/split handling and training interfaces for single and combinatorial perturbations.
Sources (3)
snap-stanford/GEARS: README.md; snap-stanford/GEARS: gears/model.py; gears: Journal full-text XML · GEARS paper Methods: Overview, Gene coexpression graph encoder, GO graph, Cross-gene effects and gene-specific decoder; gears/model.py: GEARS_Model

Limitations and conditions

The authors explicitly warn against cross-cell-type transfer, bulk-RNA assumptions and predicting combinations after training only on single perturbations.
Sources (3)
snap-stanford/GEARS: README.md; snap-stanford/GEARS: gears/model.py; gears: Journal full-text XML · GEARS paper Methods: Overview, Gene coexpression graph encoder, GO graph, Cross-gene effects and gene-specific decoder; gears/model.py: GEARS_Model

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: catalog-model-gears

Specifications

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.

Inputs, outputs and configuration
Property	Description and evidence
Model type	Graph-based perturbation-response predictor Sources (3) snap-stanford/GEARS: README.md; snap-stanford/GEARS: gears/model.py; gears: Journal full-text XML · GEARS paper Methods: Overview, Gene coexpression graph encoder, GO graph, Cross-gene effects and gene-specific decoder; gears/model.py: GEARS_Model
Architecture	Two graph encoders represent gene coexpression and Gene Ontology perturbation similarity. Perturbation embeddings are composed with gene embeddings, then a cross-gene network and gene-specific decoders predict expression changes. Sources (3) snap-stanford/GEARS: README.md; snap-stanford/GEARS: gears/model.py; gears: Journal full-text XML · GEARS paper Methods: Overview, Gene coexpression graph encoder, GO graph, Cross-gene effects and gene-specific decoder; gears/model.py: GEARS_Model
Inputs	Single-cell expression data, perturbation labels and the graph resources used by the configured model. Sources (3) snap-stanford/GEARS: README.md; snap-stanford/GEARS: gears/model.py; gears: Journal full-text XML · GEARS paper Methods: Overview, Gene coexpression graph encoder, GO graph, Cross-gene effects and gene-specific decoder; gears/model.py: GEARS_Model
Outputs	Predicted gene-expression responses and genetic-interaction analyses. Sources (3) snap-stanford/GEARS: README.md; snap-stanford/GEARS: gears/model.py; gears: Journal full-text XML · GEARS paper Methods: Overview, Gene coexpression graph encoder, GO graph, Cross-gene effects and gene-specific decoder; gears/model.py: GEARS_Model
Parameters	Configuration-dependent: gene and perturbation embedding tables grow with the selected gene/perturbation inventory, alongside graph and decoder parameters. Sources (3) snap-stanford/GEARS: README.md; snap-stanford/GEARS: gears/model.py; gears: Journal full-text XML · GEARS paper Methods: Overview, Gene coexpression graph encoder, GO graph, Cross-gene effects and gene-specific decoder; gears/model.py: GEARS_Model
Known versions	The README describes v0.1.1 updates; a specific trained checkpoint must be recorded separately. Sources (3) snap-stanford/GEARS: README.md; snap-stanford/GEARS: gears/model.py; gears: Journal full-text XML · GEARS paper Methods: Overview, Gene coexpression graph encoder, GO graph, Cross-gene effects and gene-specific decoder; gears/model.py: GEARS_Model
Training data	Fitted to the selected perturbation screen. Examples include Norman, Adamson and Dixit; the repository also lists Replogle RPE1/K562 loaders. Sources (3) snap-stanford/GEARS: README.md; snap-stanford/GEARS: gears/model.py; gears: Journal full-text XML · GEARS paper Methods: Overview, Gene coexpression graph encoder, GO graph, Cross-gene effects and gene-specific decoder; gears/model.py: GEARS_Model
Training cutoff	Inapplicable as a universal pretrained-model cutoff: GEARS fits the provided perturbation training set and builds its coexpression graph from that set. · Not applicable Sources (3) snap-stanford/GEARS: README.md; snap-stanford/GEARS: gears/model.py; gears: Journal full-text XML · GEARS paper Methods: Overview, Gene coexpression graph encoder, GO graph, Cross-gene effects and gene-specific decoder; gears/model.py: GEARS_Model
Context limits	A gene-expression vector and perturbation set over the configured gene inventory; no fixed nucleotide or amino-acid token window. Sources (3) snap-stanford/GEARS: README.md; snap-stanford/GEARS: gears/model.py; gears: Journal full-text XML · GEARS paper Methods: Overview, Gene coexpression graph encoder, GO graph, Cross-gene effects and gene-specific decoder; gears/model.py: GEARS_Model
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 sources Sources (4) snap-stanford/GEARS: README.md; snap-stanford/GEARS: gears/model.py; gears: Journal full-text XML; snap-stanford/GEARS: LICENSE · GEARS paper Methods: Overview, Gene coexpression graph encoder, GO graph, Cross-gene effects and gene-specific decoder; gears/model.py: GEARS_Model; LICENSE: licence text
Access	Official project documentation and implementation: https://github.com/snap-stanford/GEARS Sources (3) snap-stanford/GEARS: README.md; snap-stanford/GEARS: gears/model.py; gears: Journal full-text XML · GEARS paper Methods: Overview, Gene coexpression graph encoder, GO graph, Cross-gene effects and gene-specific decoder; gears/model.py: GEARS_Model
Code licence	MIT Sources snap-stanford/GEARS: 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.

60 evidence rows matching the loaded filters

Claims, original sources and review scope · Release 2026-09-23-2b89723c6dd9
Property and statement	Original source and location	Review and provenance
Diagram caption Conceptual summary of the documented data flow; optional inputs and configured downstream stages must be reported for a reproducible evaluation. Individual claims	snap-stanford/GEARS: gears/model.py Original source ↗ GEARS paper Methods: Overview, Gene coexpression graph encoder, GO graph, Cross-gene effects and gene-specific decoder; gears/model.py: GEARS_Model Shared locator for this statement’s cited sources; not a separate locator for each citation. Version: f374e43e197b295016d80395d7a54ddb81cc6769 Retrieved: 2026-09-16T19:46:20.191253+00:00	source checked automated source review · 2026-09-16 Audit 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. Field: `attributes.profile.diagram.caption` Source artifact SHA-256: `07fee864bc5020807c90ed5443ff17530116f47f257332466243c233ef2bc857` Hash scope: SHA-256 of retrieved original artifact bytes Format: original_artifact Inspected artifact
Diagram caption Conceptual summary of the documented data flow; optional inputs and configured downstream stages must be reported for a reproducible evaluation. Individual claims	snap-stanford/GEARS: README.md Original source ↗ GEARS paper Methods: Overview, Gene coexpression graph encoder, GO graph, Cross-gene effects and gene-specific decoder; gears/model.py: GEARS_Model Shared locator for this statement’s cited sources; not a separate locator for each citation. Version: f374e43e197b295016d80395d7a54ddb81cc6769 Retrieved: 2026-09-16T19:46:20.191253+00:00	source checked automated source review · 2026-09-16 Audit 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. Field: `attributes.profile.diagram.caption` Source artifact SHA-256: `fe78b7b1ede50c673b8f0d91fb2637707685c78b269ee157edb6cbc4d510d61b` Hash scope: SHA-256 of retrieved original artifact bytes Format: original_artifact Inspected artifact
Diagram caption Conceptual summary of the documented data flow; optional inputs and configured downstream stages must be reported for a reproducible evaluation. Individual claims	gears: Journal full-text XML Original source ↗ GEARS paper Methods: Overview, Gene coexpression graph encoder, GO graph, Cross-gene effects and gene-specific decoder; gears/model.py: GEARS_Model Shared locator for this statement’s cited sources; not a separate locator for each citation. Version: Retrieved page snapshot; no immutable publisher revision supplied Retrieved: 2026-09-16T20:16:14.422594+00:00	source checked automated source review · 2026-09-16 Audit 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. Field: `attributes.profile.diagram.caption` Source artifact SHA-256: `6e9a8f4a2b8ccbc9aa46c43a34701b687f71b797048adf27aae2a99d9e2dcad9` Hash scope: SHA-256 of retrieved original artifact bytes Format: original_artifact Inspected artifact
Diagram steps Perturbation screen Gene and perturbation graph representations GEARS prediction Expression response Individual claims	snap-stanford/GEARS: gears/model.py Original source ↗ GEARS paper Methods: Overview, Gene coexpression graph encoder, GO graph, Cross-gene effects and gene-specific decoder; gears/model.py: GEARS_Model Shared locator for this statement’s cited sources; not a separate locator for each citation. Version: f374e43e197b295016d80395d7a54ddb81cc6769 Retrieved: 2026-09-16T19:46:20.191253+00:00	source checked automated source review · 2026-09-16 Audit 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. Field: `attributes.profile.diagram.steps` Source artifact SHA-256: `07fee864bc5020807c90ed5443ff17530116f47f257332466243c233ef2bc857` Hash scope: SHA-256 of retrieved original artifact bytes Format: original_artifact Inspected artifact
Diagram steps Perturbation screen Gene and perturbation graph representations GEARS prediction Expression response Individual claims	snap-stanford/GEARS: README.md Original source ↗ GEARS paper Methods: Overview, Gene coexpression graph encoder, GO graph, Cross-gene effects and gene-specific decoder; gears/model.py: GEARS_Model Shared locator for this statement’s cited sources; not a separate locator for each citation. Version: f374e43e197b295016d80395d7a54ddb81cc6769 Retrieved: 2026-09-16T19:46:20.191253+00:00	source checked automated source review · 2026-09-16 Audit 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. Field: `attributes.profile.diagram.steps` Source artifact SHA-256: `fe78b7b1ede50c673b8f0d91fb2637707685c78b269ee157edb6cbc4d510d61b` Hash scope: SHA-256 of retrieved original artifact bytes Format: original_artifact Inspected artifact
Diagram steps Perturbation screen Gene and perturbation graph representations GEARS prediction Expression response Individual claims	gears: Journal full-text XML Original source ↗ GEARS paper Methods: Overview, Gene coexpression graph encoder, GO graph, Cross-gene effects and gene-specific decoder; gears/model.py: GEARS_Model Shared locator for this statement’s cited sources; not a separate locator for each citation. Version: Retrieved page snapshot; no immutable publisher revision supplied Retrieved: 2026-09-16T20:16:14.422594+00:00	source checked automated source review · 2026-09-16 Audit 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. Field: `attributes.profile.diagram.steps` Source artifact SHA-256: `6e9a8f4a2b8ccbc9aa46c43a34701b687f71b797048adf27aae2a99d9e2dcad9` Hash scope: SHA-256 of retrieved original artifact bytes Format: original_artifact Inspected artifact
Diagram title GEARS workflow Individual claims	snap-stanford/GEARS: gears/model.py Original source ↗ GEARS paper Methods: Overview, Gene coexpression graph encoder, GO graph, Cross-gene effects and gene-specific decoder; gears/model.py: GEARS_Model Shared locator for this statement’s cited sources; not a separate locator for each citation. Version: f374e43e197b295016d80395d7a54ddb81cc6769 Retrieved: 2026-09-16T19:46:20.191253+00:00	source checked automated source review · 2026-09-16 Audit 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. Field: `attributes.profile.diagram.title` Source artifact SHA-256: `07fee864bc5020807c90ed5443ff17530116f47f257332466243c233ef2bc857` Hash scope: SHA-256 of retrieved original artifact bytes Format: original_artifact Inspected artifact
Diagram title GEARS workflow Individual claims	snap-stanford/GEARS: README.md Original source ↗ GEARS paper Methods: Overview, Gene coexpression graph encoder, GO graph, Cross-gene effects and gene-specific decoder; gears/model.py: GEARS_Model Shared locator for this statement’s cited sources; not a separate locator for each citation. Version: f374e43e197b295016d80395d7a54ddb81cc6769 Retrieved: 2026-09-16T19:46:20.191253+00:00	source checked automated source review · 2026-09-16 Audit 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. Field: `attributes.profile.diagram.title` Source artifact SHA-256: `fe78b7b1ede50c673b8f0d91fb2637707685c78b269ee157edb6cbc4d510d61b` Hash scope: SHA-256 of retrieved original artifact bytes Format: original_artifact Inspected artifact
Diagram title GEARS workflow Individual claims	gears: Journal full-text XML Original source ↗ GEARS paper Methods: Overview, Gene coexpression graph encoder, GO graph, Cross-gene effects and gene-specific decoder; gears/model.py: GEARS_Model Shared locator for this statement’s cited sources; not a separate locator for each citation. Version: Retrieved page snapshot; no immutable publisher revision supplied Retrieved: 2026-09-16T20:16:14.422594+00:00	source checked automated source review · 2026-09-16 Audit 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. Field: `attributes.profile.diagram.title` Source artifact SHA-256: `6e9a8f4a2b8ccbc9aa46c43a34701b687f71b797048adf27aae2a99d9e2dcad9` Hash scope: SHA-256 of retrieved original artifact bytes Format: original_artifact Inspected artifact
Model type Graph-based perturbation-response predictor Individual claims	snap-stanford/GEARS: gears/model.py Original source ↗ GEARS paper Methods: Overview, Gene coexpression graph encoder, GO graph, Cross-gene effects and gene-specific decoder; gears/model.py: GEARS_Model Shared locator for this statement’s cited sources; not a separate locator for each citation. Version: f374e43e197b295016d80395d7a54ddb81cc6769 Retrieved: 2026-09-16T19:46:20.191253+00:00	source checked automated source review · 2026-09-16 Audit 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. Field: `attributes.profile.facts.0.value` Source artifact SHA-256: `07fee864bc5020807c90ed5443ff17530116f47f257332466243c233ef2bc857` Hash scope: SHA-256 of retrieved original artifact bytes Format: original_artifact Inspected artifact

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

Sources and history

View linked audit checks and correction history

Release 2026-09-23-2b89723c6dd9 · Record review: discovered

5 source records and release history

GEARS official resource · Original source · Not reported
snap-stanford/GEARS: README.md · Original source · f374e43e197b295016d80395d7a54ddb81cc6769
snap-stanford/GEARS: gears/model.py · Original source · f374e43e197b295016d80395d7a54ddb81cc6769
gears: Journal full-text XML · Original source · Retrieved page snapshot; no immutable publisher revision supplied
snap-stanford/GEARS: LICENSE · Original source · f374e43e197b295016d80395d7a54ddb81cc6769

Download this release

Technical metadata and extraction receipts

Stable ID: catalog-model-gears

areas: cells-tissues
method types: specialist
entity level: family
version: published implementation
reported name: GEARS
access: Public code; task-specific training data required.
method type: specialist
historical missing metadata: checkpoint revision: not_yet_extracted; training data: not_yet_extracted; licence: not_yet_extracted
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.
entity classification: review date: 2026-09-17; rationale: The cited profile describes a named learned biological predictor or representation model/family. Preserve this identity separately from task-specific fitting, individual checkpoints, pipelines and hosted access.; source ids: evidence-official-c037e3419c04936262a0; evidence-official-b5e81b2c497214b18f8a; evidence-official-f464cd407a6a9d031720; source locator: GEARS paper Methods: Overview, Gene coexpression graph encoder, GO graph, Cross-gene effects and gene-specific decoder; gears/model.py: GEARS_Model; ambiguities: None recorded

Related records

family: GEARS
alias of: GEARS
family: GEARS
subject: GEARS: family GEARS

Suggest a correction