This website uses cookies so that we can provide the best service possible. For more information, please visit our privacy policy.

sign in

CoLibri: success stories

Boehringer-Ingelheim created with our help BI-CLAIM (Boehringer Ingelheim Comprehensive Library of Accessible Innovative Molecules) covering their in-house chemistry and routinely search this proprietary space today with FTrees-FS, consisting of trillions of virtual molecules that are immediately synthesizable. They found nm inhibitors for a GPCR and proteinase project:

Identification of New Potent GPR119 Agonists by Combining Virtual Screening and Combinatorial Chemistry
B. Wellenzohn, U. Lessel, A. Beller, T. Isambert, C. Hoenke, and B. Nosse
J. Med. Chem., 2012, 55 (24), pp 11031-11041

Pfizer used CoLibri to define a variant of PGVL (Pfizer Global Virtual Library) and FTrees to search their in-house chemistry libraries. PGVL made up a space of more than 3 trillion compounds and it could be exemplified how searches find molecules which would not have been found with other similarity search methods. This proves the usefulness of the technology and the orthogonality of FTrees to other methods:

Similarity Searching and Scaffold Hopping in Synthetically Accessible Combinatorial Chemistry Spaces
M. Boehm, T.-Y. Wu, H. Claussen and C. Lemmen
J. Med. Chem., 2008, 51 (8), pp 2468-2480

Using KnowledgeSpace™ exactly as envisioned as a blueprint, Evotec employed CoLibri to shape a unique resource called ("EVOspace") based on their in-house chemistry experience and their 36 Mio compound collection ("EVOsource"). Furthermore, in collaboration, powerful KNIME®-workflows were developed, both for space-generation and -searching. These were first introduced at the KNIME® Spring Summit 2016:

Just in KNIME: Successful Process Driven Drug Discovery
M. Mazanetz
KNIME® Spring Summit, Berlin, Germany, Feb. 2016

Scientists at AstraZeneca did a systematic study about the orthogonality of different similarity search methods. Specifically it was determined across a number of data sets for a variety of targets, which method uniquely found active molecules, namely active molecules that the other methods missed. Later this study was repeated using FTrees confirming that FTrees performed best according to this benchmark.

Multifingerprint Based Similarity Searches for Targeted Class Compound Selection
T. Kogej, O. Engkvist, N. Blomberg, and S. Muresan
J. Chem. Inf. Model., 2006, 46 (3), pp 1201-1213

Using the FTrees similarity search method, Gedeon Richter found Novel Histamine H4 and Serotonin Transporter Ligands. This study also nicely confirms that FTrees is well able to jump chemical classes and find remotely similar molecules – active against the same target – with distinctly different scaffolds.

Discovery of Novel Histamine H4 and Serotonin Transporter Ligands Using the Topological Feature Tree Descriptor
R. Kiss, M. Sándor, A. Gere, É. Schmidt, G.T. Balogh, B. Kiss, L. Molnár, C. Lemmen, and G.M. Keseru
J. Chem. Inf. Model., 2012, 52 (1), pp 233-242


© 2019 BioSolveIT GmbH,   all rights reserved