metadata
license: cc0-1.0
task_categories:
- tabular-regression
- graph-ml
- text-classification
tags:
- chemistry
- biology
- medical
pretty_name: ASAP-OpenADMET LogD
size_categories:
- n<1K
configs:
- config_name: default
data_files:
- split: train
path: logd.csv
ASAP-OpenADMET LogD
ASAP_OpenADMET_LogD dataset from the ASAP Discovery-OpenADMET Antiviral Drug Discovery Challenge [1] [2] [3]. It is intended to be used through scikit-fingerprints library.
The task is to predict LogD (distribution coefficient) of molecules.
| Characteristic | Description |
|---|---|
| Tasks | 1 |
| Task type | regression |
| Total samples | 478 |
| Recommended split | time |
| Recommended metric | MAE |
References
[1] ASAP Discovery "ASAP Discovery x OpenADMET Antiviral Drug Discovery Challenge" https://polarishub.io/blog/antiviral-competition
[2] Chodera et al. "The ASAP Discovery Antiviral Drug Discovery Challenge" https://doi.org/10.26434/chemrxiv-2025-zd9mr
[3] MacDermott-Opeskin, Hugo, et al. "A computational community blind challenge on pan-coronavirus drug discovery data" J. Chem. Inf. Model. 2026, 66, 6, 3129-3149 https://doi.org/10.1021/acs.jcim.5c02106