Small Molecule Generate a low-energy bioactive conformer for a flexible drug-like molecule with RDKit’s ETKDG + MMFF pipeline — useful before docking, alignment, or shape-based screening. Dock an inhibitor into a protein crystal structure and surface the key polar contacts and hydrophobic interactions. Use a generative model to propose drug-like analogs for a covalent oncology target, optimized jointly for binding affinity, drug-likeness, and synthetic feasibility. Predict ADMET properties for a panel of approved drugs, benchmark them against DrugBank percentiles, and pinpoint the structural features driving PK liabilities. Pull compound structures and activity data straight out of a published patent — no manual structure-by-structure transcription, no SMILES typed by hand. Turn raw screening output into a docking-vs-drug-likeness scatter, ADMET liability heatmap, and radar comparison of top hits — all in a single agent loop. GPCR Dock a clinical compound set into two related GPCR subtypes and surface the structural features that drive subtype preference — a starting point for selectivity-engineered analogs. Predict and compare CNS penetration markers and metabolic stability across a competitor set — useful when triangulating PK differentiation early in a CNS program. Find recent filings around a target class, extract the lead-compound structures, and organize them with reported activities — a starting point for prior-art and freedom-to-operate work. Macrocyclic Peptide Predict how a cyclic RGD scaffold binds its receptor, identify the contacts that drive potency, and propose modifications to extend the binding signature. Extract disclosed peptide sequences and binding data from a key patent and summarize the SAR — a head-start when scoping a new peptide program.
Nucleic Acid Tx Predict the aptamer–protein complex with Boltz-2, then sweep nucleotide substitutions to find variants with stronger binding — a route to optimizing oligonucleotide therapeutics in silico. Use deep-learning splicing predictions to locate the splice regulatory elements driving exon inclusion, and identify where to target an antisense oligonucleotide for therapeutic effect. Turn position-by-nucleotide affinity data into a heatmap that highlights critical positions and beneficial mutations — quick triage for an aptamer optimization program. 
