Skip to main content
model-family passport · Review date not recorded

GEMS

Graph neural networks for potency on tough targets.

1/7Evidence fields documented
60-SECOND EVALUATION VIEW

What should a scientist know before using GEMS?

UnresolvedEvidence direction is incomplete or not yet resolved
Best suited forGeneration · Prediction
Evidence supportsPrimary links may be present, but BioAtlas does not claim a review date without a record-level timestamp.
Evidence does not establishUniversal superiority, therapeutic success, clinical utility or regulatory acceptance.
Major limitationIndependent reproducibility is limited by proprietary access.
Current registry recordVersion history not yet curated1 recorded release · Review date not recorded. A newer version is not assumed to be universally better.

What it is

Genesis's GEMS platform uses geometric and graph neural networks to predict binding, selectivity and drug-like properties, tackling targets that resist conventional methods. The company pairs its models with wet-lab cycles to advance its own pipeline.

Evidence trail

BioAtlas keeps the path from source to decision visible. A connection records provenance; it does not imply that evidence is sufficient for every context.

Sources1 connectedPrimary resources and normalized claims
Claims0 normalizedNo normalized claim yet
EntityGEMSmodel-family · Version history not yet curated
ReviewReview date not recordedReview date not claimed
ConclusionContext requiredAdd to an evaluation before operational use

Model passport

Entity typemodel-family
OrganizationGenesis Therapeutics
Model family introducedNot normalized
AccessProprietary
Commercial useVendor terms
DeploymentVendor managed
ComputeGPU or managed service
Domainschemistry
Biology → representation → computation → evidence

How GEMS represents biology

model-familychemistry

Category is navigation. These fields describe the model-specific computational transformation and deliberately override broad category defaults.

1 · Biological inputs
Molecular structures or discovery objectives
2 · Input representation
Molecular graph / tokens / 3D geometry
3 · Internal representation
Molecular representation
4 · Architecture
Chemistry model or platform
5 · Learning objective
Molecular prediction or generation
6 · Output representation
MoleculesCoordinatesScores

Biological scale

Modalities & tasks

MoleculeGenerationPredictionDocking

Registry, claims and frontier intelligence

Versioned registry

Version history not yet curated

1 version record · release year not yet normalized. Model-family identity remains separate from capability and access changes.

Explore version lineage →
Benchmark claim ledger

0 normalized claims

No task, dataset, split and metric claim has been normalized for this record yet.

Open claim intelligence →

Inputs and outputs

Inputs

Molecular structures or discovery objectives

Outputs

MoleculesScores or poses

Scientific and technical profile

Scientific principles

Graph neural networksGeometric deep learningMolecular property prediction

Technology

3D GNNsActive learningGenerative optimization
Ideas before algorithms

Scientific lineage

Explore all foundations

These are transparent concept matches—not claims that one scientist alone caused this model. Each connection is based on the model’s recorded domain, scientific principles, technical terms or an explicit lineage link.

Medicinal chemistry & pharmacology

Quantitative structure–activity relationships

Corwin Hansch

Classical QSAR established the central premise that molecular features can predict potency and guide optimization—the conceptual ancestor of modern molecular machine learning.

Matched concepts: property prediction, optimization, molecule
Medicinal chemistry & pharmacology

Rule of Five and oral drug-likeness

Christopher A. Lipinski

Generative chemistry and lead optimization routinely use drug-likeness and developability constraints inspired by this work.

Matched concepts: drug-like, molecular property
Computational intelligence

Denoising diffusion generative models

Jascha Sohl-Dickstein, Jonathan Ho and collaborators

Modern protein-backbone, molecular-pose and biomolecular-complex generators use diffusion to sample valid three-dimensional structures and designs.

Matched concepts: generative, pose
Medicinal chemistry & pharmacology

Selective toxicity and the ‘magic bullet’

Paul Ehrlich

Target selectivity, therapeutic index and mechanism-based screening remain central goals of drug discovery.

Matched concepts: target
Molecular recognition

Cooperative ligand binding

Archibald V. Hill

Dose–response curves, receptor occupancy, multisite binding and systems pharmacology still use Hill-type models.

Matched concepts: binding

Evaluation evidence

Dataset or evaluationNot yet curated
Task or metricNot yet extracted
Evidence statusNo task-specific benchmark record curated
Open source ↗

BioAtlas has not yet extracted a structured benchmark claim for this record.

Known limitations

  • Independent reproducibility is limited by proprietary access.
  • A structured benchmark claim has not yet been extracted for this record.
  • Outputs require task-specific scientific and experimental validation.

Milestones

Not normalized

Founded from Stanford ML + chemistry research.

Evidence

Raised a large Series B for the GEMS platform.