Latest
QSP
Using QSP to explain a dose-response plateau
A mechanistic model can distinguish exposure limits from biological limits and guide the next measurement.
Medicinal chemistry
A methyl group can change more than lipophilicity
Interpreting methyl substitutions through conformational preference, local interactions, and measured properties.
Target modalities
Choosing a modality starts with the biology
A practical comparison of inhibition, protein depletion, and suppression of protein synthesis.
PROTACs and molecular glues
Designing productive complexes for PROTACs and molecular glues
Productive protein recruitment depends on geometry, cellular context, and the steps after ternary-complex formation.
Drug discovery overview
Choosing a hit series that can survive the next experiment
Hit confirmation and tractability should influence a program before a potency ranking becomes a chemistry plan.
Protein-ligand modeling
Building a protein-ligand complex that can guide chemistry
Preparation, pose selection, and experimental restraints determine whether a complex is useful for design.
Target validation
A knockout phenotype is only part of target validation
Connecting human genetics, perturbation experiments, and chemical evidence before committing to a target.
AI/ML predictions
Testing whether an ML model can predict the next chemistry round
Validation should reflect the compounds, assays, and decisions the model will encounter after deployment.
Physics based modeling
Deciding which analogs deserve a free-energy calculation
Relative binding free energies are most useful when the chemical question and simulation setup are well matched.
Physicochemical properties
A solubility prediction needs a pH and a solid form
Separating ionization, hydration, and crystal packing before choosing a chemical modification.
Discovery trends
Recent discovery methods face a prospective test
Structure prediction and generative chemistry are changing project workflows; experimental follow-up determines their value.
Medicinal chemistry
Growing a fragment without losing the reason it binds
Weak binding can be a useful start when structural evidence and growth vectors support a tractable series.