Deep Learning in Pharmacology
Term: 8/2017 - 9/2019 (24 months)
Topic:
The goal the project "Deep Learning in Pharmacology" is to make drug development more efficient and to develop safe and effective drug candidates with the use of Deep Learning. Initially, the main focus of "Deep Learning in Pharmacology" is to prioritize chemical compounds in ongoing drug discovery projects and thereby increase hit rates of drug screening experiments to provide novel drug candidates. Secondly, Deep Learning models should flag chemical compounds with potentially unfavourable (e.g. toxicity-related) effects and hence focus efforts on safer drug candidates. The ultimate aim is to identify the drug targets and biological mechanisms underlying these novel drug candidates.
To achieve these goals, the objectives of the "Deep Learning in Pharmacology" project are to:
- develop accurate Deep Learning models that predict pharmacological effects of chemical compounds
- develop accurate Deep Learning models that predict toxic effects of chemical compounds
- improve the accuracy of Deep Learning models by automatically learning molecular descriptors and representations of chemical compounds
- improve the accuracy of Deep Learning models by combining public and private bioactivity data
- empower and complement the Deep Learning models by using bioassay measurements, high-content imaging (HCI), or genomic measurements as inputs
- suggest and prioritize chemical compounds for ongoing drug discovery projects
- identify novel chemical scaffolds with favourable properties for ongoing drug discovery projects
- develop Deep Learning models that accurately identify a compound's biomolecular targets