IgLM / AntiBERTa benchmarks

Antibody LM eval — paratope prediction, CDR generation, developability.

Composite

77.5

Experimental validation

Wet-lab confirmed

Stages

Hit IDDevelopmental Candidate

Modalities

biologic-mab

Task types

antibody-generationliability-prediction

Size

sequences: 600,000,000
tasks: 6

License

MIT

First release

2022

Last updated

2024-08

Official site

→ project page

Leaderboard

→ leaderboard

Dataset

→ dataset

Code / GitHub

→ repository

HuggingFace

→ HF

Paper

Generative language models for antibody design · Shuai RW, Ruffolo JA, Gray JJ · 2023 · paper · doi:10.1016/j.cels.2023.07.001 · 140 citations

Flags

none

Experts

Jeffrey J. Gray

Groups

Gray Lab (Johns Hopkins)

Hosted by

—

Related benchmarks

Observed Antibody Space (OAS), SAbDab

Rubric (7-criterion)

rigor

coverage

maintenance

adoption

quality

accessibility

industry_relevance

Notes

Moves toward true developability benchmarks.

← Back to all benchmarks

Compare:

Open comparison →