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

IgLM / AntiBERTy

The antibody-specific language models many tools build on.

4/7Evidence fields documented
60-SECOND EVALUATION VIEW

What should a scientist know before using IgLM / AntiBERTy?

SupportedEvidence supports the stated context with explicit boundaries
Best suited forGeneration · Optimization
Evidence supportsAntibody sequence evaluations: Peer-reviewed
Evidence does not establishUniversal superiority, therapeutic success, clinical utility or regulatory acceptance.
Major limitationPerformance depends on the evaluation dataset and operating conditions.
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

IgLM is a generative language model for antibody sequences enabling controllable design and humanization, while AntiBERTy provides antibody representations for downstream tasks. Together they helped establish antibody-specific language modeling and underpin tools like the IgFold structure predictor.

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.

Sources4 connectedPrimary resources and normalized claims
Claims1 normalizedAntibody sequence modelling
EntityIgLM / AntiBERTymodel-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
OrganizationJohns Hopkins (Gray Lab)
Model family introducedNot normalized
AccessOpen source
Commercial useAllowed / verify checkpoint terms
DeploymentSelf-hosted
ComputeManaged platform or GPU
Domainsantibody
Biology → representation → computation → evidence

How IgLM / AntiBERTy represents biology

model-familyantibody

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

1 · Biological inputs
Antigen, sequence or desired properties
2 · Input representation
Sequence and/or antigen geometry
3 · Internal representation
Antibody representation
4 · Architecture
Antibody model
5 · Learning objective
Antibody prediction / design
6 · Output representation
SequenceStructureScores

Biological scale

Modalities & tasks

AntibodyProteinGenerationOptimizationPrediction

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 →

Inputs and outputs

Inputs

Antigen, sequence or desired properties

Outputs

Antibody candidatesAffinity or developability estimates

Scientific and technical profile

Scientific principles

Antibody language modelingControllable generation

Technology

Autoregressive Ig LMBERT-style antibody encoder
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.

Biologics & genome engineering

Phage display and selection of binding proteins

George P. Smith & Sir Gregory P. Winter

Display-based selection created an experimental search engine for protein binders and remains a core validation partner for computational antibody design.

Matched concepts: antibody, affinity, humanization
Structural biology

Anfinsen’s dogma—the thermodynamic hypothesis

Christian B. Anfinsen

Protein structure prediction, inverse folding and generative protein design all assume that sequence strongly constrains structure and function.

Matched concepts: sequence, fold
Biologics & genome engineering

Hybridoma production of monoclonal antibodies

Georges J. F. Köhler & César Milstein

Therapeutic antibodies, diagnostic antibodies and antibody discovery platforms became scalable and reproducible.

Matched concepts: antibody, antigen
Biologics & genome engineering

Somatic gene rearrangement generates antibody diversity

Susumu Tonegawa

Antibody language models and repertoire design operate on the sequence space created by V(D)J recombination and somatic diversification.

Matched concepts: antibody, sequence
Computational intelligence

Transformer self-attention

Ashish Vaswani and colleagues

Protein, genome, molecule and single-cell foundation models use attention to learn dependencies across biological sequences and multimodal inputs.

Matched concepts: language model, sequence
Computational intelligence

Information, entropy and communication

Claude E. Shannon

Sequence modelling, cross-entropy training, language models, mutual information and representation learning all use Shannon’s framework.

Matched concepts: language model, sequence, representation

Evaluation evidence

Dataset or evaluationAntibody sequence evaluations
Task or metricAntibody language modelling
Evidence statusPeer-reviewed
Open source ↗

Task-specific evidence only; not comparable as a universal leaderboard score.

Antibody sequence modelling

Antibody sequence evaluations

Version history not yet curated · Antibody sequence evaluation sets
peer-reviewed

Peer-reviewed antibody language-modelling evaluation.

Claim caveats
  • Sequence plausibility does not guarantee affinity, specificity or developability.
  • Training-set lineage and germline distribution affect generalization.

Known limitations

  • Performance depends on the evaluation dataset and operating conditions.
  • Task-specific benchmark results should not be compared across unlike domains.
  • Outputs require task-specific scientific and experimental validation.

Milestones

Not normalized

Underpins the IgFold antibody structure predictor.

Evidence

Openly released for research.