piotr
Update README.md
7372370 verified
metadata
license: unknown
task_categories:
  - tabular-classification
  - graph-ml
  - text-classification
tags:
  - chemistry
  - biology
  - medical
pretty_name: TDC Carcinogens Lagunin
size_categories:
  - n<1K
configs:
  - config_name: default
    data_files:
      - split: train
        path: tdc_carcinogens_lagunin.csv

TDC Carcinogens Lagunin

Carcinogens Lagunin dataset dataset [1] [2], part of TDC [3] benchmark. It is intended to be used through scikit-fingerprints library.

The task is to predict whether the drug is a carcinogen.

Characteristic Description
Tasks 1
Task type classification
Total samples 280
Recommended split scaffold
Recommended metric AUROC

References

[1] Lagunin, Alexey, et al. Computer-Aided Prediction of Rodent Carcinogenicity by PASS and CISOC-PSCT" QSAR & Combinatorial Science 28.8 (2009): 806-810 https://doi.org/10.1002/qsar.200860192

[2] Cheng, Feixiong, et al. "admetSAR: A Comprehensive Source and Free Tool for Assessment of Chemical ADMET Properties" Journal of Chemical Information and Modeling 52.11 (2012): 2840-2847 https://doi.org/10.1021/ci300367a

[3] Huang, Kexin, et al. "Therapeutics Data Commons: Machine Learning Datasets and Tasks for Drug Discovery and Development" Proceedings of Neural Information Processing Systems, NeurIPS Datasets and Benchmarks, 2021 https://openreview.net/forum?id=8nvgnORnoWr