Prof. Andrea Volkamer, Saarland University, Saarbrücken, Germany

Open-Source Developments for Structure-Based Kinase-Centric Drug Design

Start Date 09.04.2024 - 16:00
Event End 09.04.2024 - 17:00
Location ETH Zürich, Hönggerberg
ETH Zürich, Hönggerberg, HCI J4
Register

Human protein kinases play a significant role in numerous diseases, making them a crucial protein family for targeted therapy. To date there are over 6,000 human kinase structures in the PDB and around 70 small molecule kinase inhibitors available. However, challenges such as drug promiscuity, resistance, and unexplored kinase territory persist.In this presentation, we will provide an overview of methods that leverage openly available kinase data to generate new insights and facilitate community engagement, with a focus on machine learning (ML) tasks. Utilizing the TeachOpenCADD platform [1,2], we will demonstrate how diverse computer-aided drug design (CADD) tasks can be orchestrated for individual kinases. Additionally, we will introduce freely available tools to support kinase research, including: (i) KinFragLib for fragment-based kinase inhibitor design [3], (ii) KiSSim – a KLIFS-based kinase structural similarity fingerprint [4], and (iii) a pipeline for assessing kinase similarity from various data perspectives [5].

Lastly, we will present ongoing projects in structure-informed ML for kinase inhibitor design and affinity prediction across kinases [6,7], employing deep learning techniques.