Exploring the Applications of Chemical Space Networks in Molecular Library Design and Drug Design

Goal: This project aims at the exploration of chemical space from a network-centric perspective. We propose to investigate molecular similarity networks using different families of molecular descriptors and similarity measures. The application of various network measures to the design of diversity oriented molecular libraries is a major goal of this project. We have shown that exploration of protein binding site similarity networks hold great potential for predicting cross reactivity of drugs.
Chemical transformation networks have thus far been relatively unexplored from a network perspective. We propose to make significant inroads in this direction using mathematical network measures, comparing the characteristics of these networks to those of molecular interaction and similarity networks. Finally, we intend to apply the outcome of our findings to generate two molecular libraries - the first via diversity oriented synthesis followed by biological evaluation against various phenotypes, and the second a target-based library against α-glucosidase.

Objectives: Comparison of molecular networks constructed from whole molecule and fragment-based, 2-D and 3-D similarity measures;

• Exploration of protein binding site similarity networks for predicting cross reactivity of drugs;
• Comparison of network measures for molecular interaction, transformation and similarity networks;
• Molecular library design, synthesis and screening based on network measures: a) Design, synthesis and phenotypic evaluation of a DOS library; b) A target based library against α-glucosidase.

The project cited above has been approved on 30 Aug.2017 by Science and Engineering Research Board (SERB) for funding, for a period of 3 years.