SPRINT
SPRINT is a fast, accurate, and scalable deep learning framework for virtual screening of thousands of molecules.
SPRINT is a vector-based deep learning framework designed to accelerate drug discovery by enabling massive-scale virtual screening of drug-target interactions (DTIs). It utilizes structure-aware protein language models and a self-attention architecture to learn a co-embedding space for both small molecules and protein targets. This approach allows for extremely fast and accurate prediction of interactions, capable of screening thousands of molecules in minutes. SPRINT achieves state-of-the-art results on virtual screening and DTI classification benchmarks, making it a powerful tool for identifying new drug candidates and predicting off-target effects.
Estimated resource cost: 0.000527
Categories: Ligand/Drug Design & Screening, Molecular Docking & Interactions
Tags: Proteins Small Molecules Virtual Screening
Key Capabilities
- Ultra-fast virtual screening enabling pan-proteome-scale DTI predictions.
- Achieves state-of-the-art performance on DTI classification, virtual screening, and binding affinity benchmarks.
- Employs structure-aware protein language models and a multi-head attention pooling scheme for improved accuracy.
Runtime Statistics
| Metric | Value |
|---|---|
| runtime_mean | 43 |
| runtime_median | 38 |
| runtime_std | 32 |
| runtime_90th_percentile | 63 |
| runtime_max | 341 |
Similar Tools
- AF2Bind
- AutoDock Vina (smina)
- DiffDock-L
- DynamicBind
Ready to submit your job?
Review your configuration, then confirm the estimated credit cost before you run the job. Note that credit estimates are not guaranteed and runtime can vary depending on inputs and settings.
Estimated Credits: 0.000527
Invite-only, limited-time access. Please contact ztang@getantibody.com.
Invite-only, limited-time access. Please contact ztang@getantibody.com.