COMET: A Machine-Learning Framework Integrating Ligand-Based and Target-Based Algorithms for Elucidating Drug Targets
Elucidation of the potential molecular targets of a bioactive compound, a process known as target-fishing, is a critical task in drug discovery. Computational methods can efficiently narrow down the candidate targets for subsequent experimental validation. We have developed a computational target-fi...
Saved in:
| Main Author: | |
|---|---|
| Other Authors: | , , , , , , , , , , |
| Published: |
2025
|
| Subjects: | |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Summary: | Elucidation of the potential molecular targets of a bioactive compound, a process known as target-fishing, is a critical task in drug discovery. Computational methods can efficiently narrow down the candidate targets for subsequent experimental validation. We have developed a computational target-fishing method, termed COMET, which integrates ligand-based similarity scores with target-based binding scores into a random forest algorithm for target ranking. COMET leverages carefully curated data sets encompassing 2685 human targets of therapeutic relevance and 990,944 ligand-target interaction pairs. Moreover, its modular framework allows convenient future upgrades. In a head-to-head comparison with seven other methods, COMET exhibited consistent top 100 recalls and top 15 hit rates of 70∼80% on two large test sets, showing clear advantage over the rivals. Moreover, the optimal balance between speed and accuracy renders COMET highly suitable for practical applications. COMET has been implemented for free trials on the PDBbind+ web server (http://www.pdbbind-plus.org/comet), and has already completed over 2700 jobs submitted by users worldwide. |
|---|