Dynamic learning to generate better compounds

Combine your expert chemistry knowledge and the exploratory power of generative methods to identify optimal compounds faster.

The Inspyra module in StarDrop

As you work, Inspyra generates new compound ideas in the background while learning from your feedback. Its Inference Engine learns your preferences across multiple properties simultaneously, considering the relationships between them, as well as your preferences for chemical fragments. Your responses guide further suggestions towards the property balance and chemistry you want to pursue.

Inspyra’s generative chemistry methods enable you to

Try Inspyra

Generate new compound ideas

Expand your compound collections or explore new analogues within a chemical series.

Explore vast regions of chemical space

Discover alternative substitutions and structural motifs, with suggestions guided by your feedback.

Discover optimisation strategies

Assess ideas across multiple properties to identify modifications that offer a promising balance for your project.

Better compounds in just a few clicks

Inspyra can be used in two modes:

Inspyra Panel: While working in StarDrop, new compound ideas are suggested in a panel below your data set. Coloured borders show how likely Inspyra predicts you are to like an idea or add it to your data set, and property plots let you compare each idea’s calculated properties with the others. When you like, dislike, ignore or add an idea, Inspyra will respond to your feedback, tailoring further suggestions to your preferences. You can also apply one action to all displayed ideas, resize the panel, choose how many ideas appear, and adjust or pause the refresh rate to review at your own pace.

Inspyra Designer: To explore modifications to a single compound, simply circle the region you want to replace and Inspyra will suggest alternative fragments, colour-coded by how likely it predicts you are to favour them, with property plots so you can compare their calculated properties. As you give feedback, Inspyra will refine its suggestions. You can then choose a replacement to keep designing, or combine replacements across regions to enumerate a small virtual library.

Generative chemistry resources

On-demand

AI in early drug discovery: from promise to practice

In this webinar, Jeff Blaney (Senior Director of Discovery Chemistry, Genentech), Darren Green (Head of Cheminformatics & Data Science, GlaxoSmithKline), Julian Levell (Head of Discovery, New Equilibrium Biosciences), Matthew Segall (CEO, Optibrium) discuss the state of AI in early drug discovery from hit to preclinical candidate and share their experiences with and expectations of AI, including predictive modelling, synthesis prediction, and generative chemistry. Hear about the successes of AI drug discovery and an outlook on what AI needs to achieve to really transform the industry.​

AI-guided discovery

Would you like to try generative chemistry to help your projects?

Inspyra is available as optional module for StarDrop. To try StarDrop please complete the form and a member of the team will get in touch to understand your needs.