NucleoBench
Open-source benchmark for nucleic acid sequence design algorithms. Evaluates 9 design algorithms across 16 diverse biological tasks (gene expression control, transcription factor binding). Over 400,000 experiments. Led to development of AdaBeam hybrid algorithm.
Composite
88.9
Experimental validation
Retrospective
Stages
Hit IDLead ID / ADMET
Modalities
dnarnanucleic_acid
Task types
sequence_designalgorithm_comparison
Size
tasks: 16
algorithms: 9
experiments: 400,000
algorithms: 9
experiments: 400,000
License
Apache-2.0
First release
2025-06
Last updated
2025-09
Official site
Leaderboard
→ leaderboard
Dataset
→ dataset
Code / GitHub
HuggingFace
→ HF
Paper
NucleoBench: A Large-Scale Benchmark of Neural Nucleic Acid Design Algorithms · · 2025 · paper · doi:10.1101/2025.06.20.660785 · 12 citations
Flags
gene_therapyrna_therapeuticsgoogle_research
Experts
—
Groups
—
Hosted by
—
Related benchmarks
Rubric (7-criterion)
rigor
5
coverage
4
maintenance
4
adoption
4
quality
5
accessibility
5
industry_relevance
4
Notes
First comprehensive benchmark for nucleic acid design. Google Research + Move37 Labs collaboration. Critical for CRISPR therapies, mRNA vaccines, and gene therapy design. AdaBeam outperforms on 11/16 tasks.