research

  • SECS: Stochastic Exploration of Chemical Space
    • Traditional virtual screening in drug discovery relies on a pre-compiled compound library (~10E9 compounds) which only covers the fraction of the chemical space (10E60 compounds).
    • Previous library-free virtual screening methods (genetic algorithem, AI) still rely on a pre-compiled fragment library or pre-trained models.
    • SECS sequentially applies the structure perturbation to sample the chemical space, analogous to trajectories in molecular dynamics, and generate novel chemical structures to be used for virtual screening.
    • Preliminary implementation can be found at https://github.com/wryu94/stochastic_exploration_of_chemical_space
  • Molecular dynamics and enhanced sampling simulation of protein-ligand complexes
    • Molecular dynamics and enhanced sampling are valuable tools to study atomistic details of pharmaceutical systems of interest:
      • many cases in which the atomistic mechanism of action for approved drugs are not known
      • or specific binding site
      • computational characterization of these protein-ligand systems can be used for lead optimization / catalyze further drug discovery work
    • Systems we are interested in:
      • HIV capsid + lenacapavir:
        • Ligand conformational space
        • Lead optimization for mutation resistance
      • HCMV terminase + letermovir:
        • Binding site identification
        • Structural modeling
      • HCMV terminase + tomeglovir:
        • Binding/unbinding free energy surface and pathway
        • Mechanism of action
  • CADD (computer aided drug discovery) collaborations
    • The Ryu Lab welcomes collaboration with medicinal chemistry, organic chemistry, and biology research groups on drug discovery projects.
    • We bring expertise in CADD methods such as:
      • Virtual screening
      • Structure prediction
      • Alchemical free energy calculations
      • Cheminformatics
      • Molecular dynamics and enhanced sampling simulation
    • At various contexts in pre-clinical drug discovery campaign such as:
      • Hit identification and generation
      • Target identification
      • Lead optimization
    • With technical proficiency in:
      • Molecular modeling software
      • Python programming