Cdk Workshop 2009

Small molecules and reaction mechanism Rewiring the enzyme space Syed Asad Rahman E-mail: [email protected]

04/21/2009

CDKWS- 2009

EBI is an Outstation of the European Molecular Biology Laboratory. 

Overview • Understanding Metabolic Networks • Shortest path based conversion of substrate to product • Challenges and some meaningful solution • Pathway Hunter Tool (PHT)

• Small Molecule similarity and its impact • Fingerprint vs Maximum Common Subgraph • SMSD (Small Molecule Subgraph Detector)

• EC classification • • • •

Do the represent the bonds of Life or are they like computer IP address? A step towards reaction homologues and its impact on EC classification Transition state analogues ERI (Enzyme Reaction Information)-Toolkit

Enzyme and reaction 1

3

5

1

1

E1

E2 2

2

3

5

2

FINDING SHORTEST PATH BIOCHEMICAL NETWORKS

Automated metabolic network reconstruction

This figure exemplifies the problem of finding a valid shortest path in the biochemical network

Rahman, S. A. et al. Bioinformatics 2005 21:1189-1193; doi:10.1093/bioinformatics/bti116 Copyright restrictions may apply.

Mapping Reactant and Product Understanding patterns in reactions

Rahman, S. A. et al. Bioinformatics 2005 21:1189-1193; doi:10.1093/bioinformatics/bti116

Pathway Hunter Tool (PHT) – A Platform for Metabolic Network Analysis and Potential Drug Targeting (PhD Thesis 2006) http://kups.ub.uni-koeln.de/volltexte/2008/2450/

Copyright restrictions may apply.

Connectivity Analysis: Metabolic Centric View

Google: Pathway Hunter Tool (PHT)

Tracing Atom(s) in the Pathway

• ARM (http://www.metabolome.jp) • PHT (http://pht.tu-bs.de/PHT/)

SIMILARITY SEARCH

Small Molecule Similarity Reaction Similarity

Matching Chemical Moieties • Index of pattern matching • Similarity (i.e. Tanimoto co-efficient) • Distance (i.e. Euclidean )

• Finding similar molecules (i.e. substructure) • MCS (Maximum common subgraph) • + Very good for substructure search • + Pinpoints the common patterns • - Computation exhaustive • Fingerprints (Descriptors of chemical moieties) • + Fast and descriptor depended • + Very good for reporting overall similarity and size of the molecule • - Can’t map the matched patterns on the query molecule

Maximum Common Subgraph (MCS)

Few words about CDK MCS • Its very fast when two molecules are very similar! Except • MICS (Maximum Common Induced Edges), hence treats cyclopropane and Isobutane as similar molecules

• Rings structures differ at non-ring sites

• Ranking the MCS solutions

Introducing a new Maximum Common Subgraph (MCS) calculation software

SMALL MOLECULE SUBGRAPH DETECTOR (SMSD) http://www.ebi.ac.uk/thornton-srv/software/SMSD/ OR Google: SMSD + EBI http://www.ebi.ac.uk/thornton-srv/software/SMSD/cdkws2009.zip

MCS vs. SMSD

Fingerprint vs. SMSD (~50,000 pairwise comparison)

Tanimoto score =

~ 0.50 as 99 percentile of the data is covered ~ 0.77 as 99.9 percentile of the data is covered

Acknowledgement • Franz Fenninger • Dr. Lorenzo Baldacci • Dr. Gemma L. Holliday • Dr. Kazuto Yamazaki • Dr. Matthew Bashton

Prof. Janet M. Thornton

Prof. R. Schrader

• Special thanks to CDK development team • • •

Chemistry Development Kit (CDK) Chemaxon Open Babel

Prof. D. Schomburg

THANK YOU!

Syed Asad Rahman E-mail: [email protected] Skype: s9asad