Background
In cancer treatment, developing highly selective small molecule inhibitors is crucial for enhancing efficacy and reducing side effects. These inhibitors precisely target specific proteins, minimizing damage too normal cells and improving drug safety
Challenge
- In a competitive market, new drugs must demonstrate superior effficacy and safety.
- Minor variations in residue subtypes of different target proteins make it challenging to accurately pinpoint selectivity sources.
Key Results
- The XFEP predictive method showed a high correlation of 0.69 with experimental results for selectivity predictions of 20 molecules across 5 subtypes.
- Using an interpretable model, we identified key amino acid residues affecting selectivity, significantly improving molecular selectivity through two rounds of targeted modifications.
