FLIP
Fitness landscape inference benchmarks with realistic train/test splits (AAV, GB1, Meltome, SCL, Bind).
Composite score: 80.8
Rubric (1–5 per criterion)
rigor
5/5
coverage
3/5
maintenance
3/5
adoption
4/5
quality
5/5
accessibility
5/5
industry_relevance
3/5
Metadata
Stages
Target IDDevelopmental Candidate
Modalities
protein-general
Task types
fitness-prediction
License
CC-BY 4.0
First release
2021-12
Last updated
2024-05
Flags
none
Size & scope
- landscapes: 5
- splits: 15
Primary paper
Title
FLIP: Benchmark tasks in fitness landscape inference for proteins
Authors
Dallago C, Mou J, Johnston KE, et al.
Year
2021
DOI / arXiv
Citations
120
Links
- Official site: https://benchmark.protein.properties/
- GitHub: https://github.com/J-SNACKKB/FLIP
- Leaderboard: N/A
Hosted by (initiatives)
Experts (primary authors / maintainers)
Groups (host labs / companies / consortia)
Related benchmarks
Notes (honest caveats)
Complements ProteinGym (smaller but carefully designed splits).