KinomeX is an online platform to predict kinome-wide polypharmacology effect of small molecules based solely on their chemical structures. The prediction is made by a multi-task deep neural network (MTDNN) model built with over 170,000 bioactivity datapoints composed of 391 kinases and ~32,000 compounds. MTDNN implements a solution to the problem with multiple and relevant tasks, here corresponding to the bioactivity prediction against a spectrum of kinases. Extensive computational and experimental validations have been performed. Overall, KinomeX enables users to create a comprehensive kinome interaction network for designing novel chemical modulators, and is of practical value on exploring the previously less studied or untargeted kinases.

The dataset used by the MTDNN model can be downloaded here. The molecular information can be found here. The details of the model are described here. The paper of KinomeX can be found here.

References:

Chartier M, Chénard T, Barker J, et al. Kinome Render: a stand-alone and web-accessible tool to annotate the human protein kinome tree[J]. PeerJ, 2013, 1: e126.
Graczyk P P. Gini coefficient: a new way to express selectivity of kinase inhibitors against a family of kinases[J]. Journal of medicinal chemistry, 2007, 50(23): 5773-5779.
Karaman M W, Herrgard S, Treiber D K, et al. A quantitative analysis of kinase inhibitor selectivity[J]. Nature biotechnology, 2008, 26(1): 127.
Bienfait B, Ertl P. JSME: a free molecule editor in JavaScript[J]. Journal of cheminformatics, 2013, 5(1): 24.
Illustration reproduced courtesy of Cell Signaling Technology, Inc. (www.cellsignal.com)