Graph Neural Network based Virtual Screen for Inotropic Compounds

How to apply

Who to contact?	sinz@cs.uni-goettingen.de
Email subject	Start with "[ML Master]"
What to include?	All applications What you are studying (e.g. "Master in Applied Data Science") Statement of motivation (max 300 words) outlining your personal interest in our research areas. A short list of which relevant skills you have and which you would like to acquire. A list of courses you have taken and corresponding grades Known programming languages and experience Possible start date If you apply with your own project idea A concise (max 300 words) “research proposal” on a project you would be interested working on. This is not binding but helps to get the conversation started.

Graph Neural Networks (GNN) are uniquely suited to represent functional compound classes in ligand based virtual screening (https://doi.org/10.1186/s13321-020-00479-8). We will setup and train a GNN on known inotrope compounds and screen i) the drug repurposing hub and ii) an in-house library for novel drug candidates that will then be phenotypically screened on our EHM based inotrope screening platform.

Background

In a recent seminal paper (https://doi.org/10.1016/j.cell.2020.01.021) Stokes and Collogues have outlined a generic approach on how to device and a generic graph neural network on a subset of chemical compounds with a given biological activity and screen large libraries of drug-like compounds for candidates with similar, in their case bactericidal, activity.

Project Description

The student will setup chemical a fingerprints database using extended connectivity fingerprints (ECFPs) and SMILES representations from existing drugs generated using OpenChem, a PyTorch based toolkit for computational chemistry. Drugs will be classified into cardioactive and cardio-innactive compounds and machine learning methods of different complxity (RNN, GNN) will be compared.

Chemistry department will provide a virtual library of undocumented drug-like compounds accessible by their novel synthesis methods. The trained NN is used to identify promising candidates that will be synthesised and tested for inotropy and chronotropy.

Offer/Profile

We offer hands-on master project which combines deep learning and computational chemistry and merges the unique on-site expertise of CIDAS, chemistry, and pharmacology. The candidate should have experience in data science and python, basic chemistry would be an asset.

Further Contact

http://www.pharmacology.med.uni-goettingen.de/RG_Meyer.html

• Dr. Tim Meyer, tim.meyer@med.uni-goettingen.de
• Dr. Svenja Warratz, swarrat@gwdg.de