Large-scale comparison of machine learning methods for drug target prediction on ChEMBL
Additional Information on dataset creation and code: http://ml.jku.at/research/lsc/index.html
Dataset Download: http://ml.jku.at/research/lsc/mydata.html
- maybe most interesting to you (in case you are working in Python) is the section
dataPythonReduced(reduced ⇒ because it is the data on which finally, the benchmark for Deep Learning was performed; everything that is unimportant to the Deep Learning code skipped) - load
dataPythonReduced-data (for the FNN case) into Python: https://github.com/ml-jku/lsc/blob/master/pythonCode/apred/loadData.py - data download script: https://github.com/ml-jku/lsc/blob/master/download.sh (maybe you need only parts out of it, but maybe simpler to use than downloading by browser)
Update Feb 2020:
- Consider that the source codes are based on an older version of TF1.X and are not adapted to the eager execution mode of TF2.
- A technical appendix and a partial reanalysis is available here. To obtain the results there, change "normalizeLocalDense" to "True" even for sparse features (since for simplicity reasons the matrix in the reference code is converted to a dense one).
BibTeX:
@Article{bib:Mayr2018,
author="Mayr, Andreas and Klambauer, G{\"u}nter and Unterthiner, Thomas and Steijaert, Marvin and Wegner, J{\"o}rg K. and Ceulemans, Hugo and Clevert, Djork-Arn{\'e} and Hochreiter, Sepp",
title={{Large-scale comparison of machine learning methods for drug target prediction on ChEMBL}},
journal="Chem. Sci.",
year="2018",
volume="9",
issue="24",
pages="5441-5451"
}