1. Background: This PhD project aims to develop and experimentally validate a novel AI tool for de novo drug designing. Developing new medicines requires considering extremely large numbers of molecules to identify suitable candidates for further investigation as therapeutics. However, traditional medicial chemistry approaches allow to explore only a small portion of the 1e+63 drug-like molecules predicted to be synthetically accessible. This project will integrate advanced machine learning techniques with molecular modelling and experimental biophysics to de novo designing compounds that bind with high affinity to selected pockets on any macromolecule of interest.
2.Novelty & Importance: This PhD project represents a novel integration of statistical search methods, explainable AI, and structural biology for molecular design. It advances the frontier of AI-driven drug discovery and promises a cost-effective, rapid, and chemically original alternative to conventional drug discovery pipelines. By lowering technical and financial barriers, the platform will enable academic and industrial researchers to tackle urgent therapeutic challenges, including neglected or hard-to-drug targets. Ultimately, this project will have a significant international impact by demonstrating the feasibility of targeting biomarkers and therapeutic avenues that have previously resisted classical approaches.
3.Aims & Objectives: This PhD project aims to develop and experimentally validate a novel AI tool: DrugSynthMC.v2 – an upgraded version of DrugSynthMC fine-tuned for the de novo generation of customized drugs with high affinity for macromolecules of interest.
The specific objectives are to:
1. Develop DrugSynthMC.v2.
2. Validate DrugSynthMC.v2 in silico.
3. Incorporate statistical and chemical compound optimization strategies.
4. Test DrugSynthMC.v2 on a non-conventional drug target.
5. Experimentally validate DrugSynthMC.v2.

