CNS BBB Library Chem Div Inc Knowledge Mining
CNS BBB Library Chem. Div Inc.
Knowledge Mining Sources Pub. Chem Database Thomson Integrity Sci. Finder Protein Data Bank Scientific publication & Patents (2010 -2015) 2
Knowledge Progression & Selection Reference Compounds > 20 K ► 2 D-Topological analogues ► Kohonen-based modeling ► BBB profiling ► 2 D-clustering procedure ► isosteric and bioisosteric morphing ► singeltones assignment FOCUS ON NOVEL CHEMISTY CNS-active agents claimed since 2010 CNS Library >17 K ► >600 Templates Design ► Focus on novel chemistry ► Med. Chem Filters 3
Reference Compounds (Examples) Disclosed since 2010 4
Methodology for Compound Selection Ref cmpds: >600 molecules with exp. determined BBB permeability 5 Kohonen-based in silico modeling BBB Permeability av. classification power: 87% Descriptors: MW, Log. P, HBA, HBD, Sp 3, Zagreb, PSA, Ss GREEN – BBB(+) ref cmpds BLUE – BBB(-) ref cmpds Points – compounds from CNS library. More than 90% are localized in BBB(+) areas or in proximity to the BBB(+)associated neurons of high priority
Descriptor Distributions 6
Validation of CNS Library Chem. Div’s Design & Activity (Pub. Med Examples) 7
CNS Library Statistics Diversity in Heterocycles: 1085 unique heterocycles Clusters (Tanimoto metrics, min cmpds per cluster – 10, similarity threshold 0. 6): 397 Singeltones: 11245 Av. structures in cluster: 15 Screens: 11938 Diversity: 0. 85 Basic aliphatic nitrogen: 42. 5% cmpds Diversity Plot 8
Examples of Scaffolds 11
Examples of Compounds 12
11 Thank you
- Slides: 11