This document discusses the potential for open source and freeware software to replace proprietary tools in chemical biology applications. It describes how open source empowers users to pursue novel enhancements according to their needs and timelines, without being constrained by proprietary software developers. However, open source must still be able to sustain effective, accurate and efficient science. The document then provides examples of specific open source tools that can successfully perform tasks like acquiring and managing data, molecular structure prediction and characterization, protein structure prediction, structure-based design, QSAR modeling, and toxicology profiling. These tools are presented as viable alternatives to proprietary software for various chemical biology and drug discovery applications.

1. Informatics for All
The Open Source and Freeware
Revolution in Chemical Biology
Applications to Selected SRI
Projects

2. Proprietary software: black box model
● Exclusive code control
● Exclusive development / customization control
User-initiated customization must
await proprietor implementation
Or
reinvent all requisite wheels

3. Open source empowers users to pursue novel
enhancements according to their needs and their
timelines!

4. Open source = philosophically good
But in practice:
can you replace well established
proprietary tools with open source
and still sustain
● Effective
● Accurate
● Efficient
science?

5. Sometimes it helps to
Yes! have a guinea pig
(mostly)
= me
July – Oct.
16 distinct
projects for
8 clients
98+% open
source

6. Synthesis & Intellectual Assay
Procurement Property Development
WWW
Meta Data Omics Meta
Data Data
Chemical
Structures Screening
Data
Target Discovery
Scope
Structure-Based Design
SAR, ADME, Tox, PK

12. Meta Data Omics Meta
Data Data
Chemical
Structures Screening
Data
Acquire
& manage
data

13. Chemical specification, drawing & editing:
Marvin (http://www.chemaxon.com/products/marvin/) functionality approaching
that of ChemDraw; good drawing options; can embed into office documents

14. Enumerate combinatorial libraries
SmiLib (http://gecco.org.chemie.uni-frankfurt.de/smilib/) Efficient and flexible
Marvin (http://www.chemaxon.com) Intuitive but slower

15. Molecular Structure Conversion
Molconverter (http://www.chemaxon.com/products/marvin/molconverter/) Fast
OpenBabel (http://www.openbabel.org) Excellent functionality but slow

16. Store / analyze libraries and screening data
Screening Assistant SA2 (http://sa2.sourceforge.net/) Powerful, enterprise-like
software: capable of handling internal data management for serious operations

17. WWW
Need Caveat:
External logged query
Knowledgebases = disclosure!

18. Chemical Data / Meta Data
ChemSpider (http://www.chemspider.com/) structure, literature, suppliers
PubChem (http://pubchem.ncbi.nlm.nih.gov/) structure, screening data
SureChem (https://surechem.com/) patent searches

19. ADME/Tox profiling; target identification
PASS (http://www.pharmaexpert.ru/passonline/index.php) only offered online,
free, and surprisingly accurate predictions on 300+ endpoints

20. ADME profiling
iLab2 (https://ilab.acdlabs.com/iLab2/) good range of ADME endpoints,
online only, one compound at a time

21. Target Discovery
Need Structure-Based Design
Modeling,
Informatics SAR, ADME, Tox, PK

22. Molecular Structure Prediction / Characterization
Avogadro (http://avogadro.openmolecules.net/)
Great builder; good graphics; built in molecular mechanics; hooks to free quantum codes

23. Molecular Structure Prediction / Characterization
VMD (http://www.ks.uiuc.edu/Research/vmd/)
Good graphics; excellent analytical tools; hooks to NAMD (molecular dynamics)

24. Molecular Structure Prediction / Characterization
PyMol (http://www.openpymol.org)
Great graphics; Decent builder, good analytical tools

25. Protein Structure Prediction
SwissModel (http://swissmodel.expasy.org/) good control, must have close homolog
Modeller (http://salilab.org/modeller/) use this for optimal control and efficient relaxation

26. Structure Based Design
PyRx / AutoDock (http://pyrx.sourceforge.net/) easy to use; good predictions
Surflex (http://www.jainlab.org/contact.html) fast; accurate; no free interface

27. QSAR: Descriptors
CDK (http://rguha.net/code/java/cdkdesc.html)
Good descriptor selection, easy to use
SA2 (http://sa2.sourceforge.net/)
Better descriptor selection, but harder to navigate

28. QSAR: modeling
BuildQSAR (http://profanderson.net/files/buildqsar.php)
Fast, flexible, easy to use

29. Toxicology profiling
ToxTree (http://toxtree.sourceforge.net/) fast, easy to use, clear logic,
good array of toxicological endpoints

30. Synthesis & Assay
Procurement Development
Target Discovery
Structure-Based Design
Information
Flow SAR, ADME, Tox, PK

31. Workflows (i.e., seamless process integration)

33. That's enough for now .....
Thank you!
Any questions?