ML Follow-up for DEL:

Identify real binders,
Visualise all pockets,
Deconvolute binding sites,
Map surface sites.

Pocket map generation on a success fee basis.

About PocketMapper

PocketMapper is our most powerful DEL follow-up system. It is a machine learning pipeline that rigorously classifies structures as signal or noise, and removes all structures identified as noise. Information from the surviving compounds is then used to create an atomic-resolution map of the binding site or binding sites.

The system is extremely robust, and we are pleased to offer it on a fee-per-success basis: if the de-noised compounds do not converge to form a pocket map, no fee is payable.

PocketMapper Reveals:

1. The number of pockets being bound by DEL hits.

2. The exact shape of each binding pocket when fully filled, including shallow or allosteric pockets.

3. The size and relative flexibility of each part of each pocket.

4. Locations within each pocket that strongly prefer hydrogen bond donors/acceptors, or hydrophobes.

5. Whether the proximal end of the linker is involved in binding.

6. The identities of a great variety of binding modes.

7. The binding modes of surface-binding pockets (e.g. dimerisation domains or allosteric regions).

PocketMapper greatly de-risks hit identification, accelerates lead optimisation, provides diverse additional paths to effective candidate chemistry, and renders a significant amount of lab work unnecessary.

About DeepDELve

DeepDELve is a highly-optimised deep learning pipeline for virtual screening. It is used as an extension to PocketMapper, allowing fast access to novel, diverse, lead-like binders that can be delivered within several weeks.

DeepDELve enables expansion out of the DEL space and into the billions of compounds easily accessible through our catalogue partners.

24
hours
To train models and search all commercial catalogues (including Enamine REAL lead-like space).
79%
hit-rate
Highest hit-rate so far (one compound ordered per scaffold, hits defined as <10µM).
90%
lower
Cost per confirmed lead, compared to resynthesis.

About ShapeRetriever

ShapeRetriever is our shape-screening pipeline, specifically optimised to use PocketMapper maps as a query. It screens into billions of our catalogue partners' high quality lead-like compounds to deliver novel chemical matter with reduced synthesis burden.

Partners

In addition to large, medium, and small pharmaceutical companies currently undisclosed.

News

5 November 2021
Cambridge Molecular and HitGen Announce Strategic Partnership
All News
background picture

Contact Us

general@cambridgemolecular.ai

Cambridge Molecular Ltd
The Bradfield Centre
184 Cambridge Science Park
Cambridge CB4 0GA
England

© 2019 Cambridge Molecular Ltd