ChemServerLigand design: molecule evolution and synthesis planning
Start from a hit, let ChemServer mutate it over generations and keep the most drug like variants. Then work out how to make the best one from building blocks you already have.
Molecule evolution
A genetic algorithm built on RDKit. From a seed SMILES it makes structural mutations, scores every new molecule and lets the best ones go on to the next generation.
- Up to 50 generations, with a population cap so runs do not grow out of control.
- Fitness from drug likeness descriptors and synthetic accessibility scores.
- The global top molecules come back with 2D images, ready to send to Batch Docking in one click.
More than 1,285,000 mutations have been evaluated on ChemServer so far.
Synthesis planner
Give a target SMILES and, if you like, a list of chemicals you have in the lab. The planner deconstructs the target step by step with reaction templates for common named reactions, then lays out the forward route with reagents and conditions, marking the starting materials you already own.
- Synthesis depth from 1 to 20 steps.
- Adjustable search breadth: wider searches find more routes but take longer.
A design loop
- Dock a set of candidates and pick a hit.
- Evolve it into a few hundred variants.
- Dock the top variants against the same pocket.
- Plan the synthesis of the best one.