Towards a comprehensive computational platform for next generation drug development – A Russian‐German joint venture

1. Towards a comprehensive
computational platform for next
generation drug development –
A Russian‐German joint venture
Edgar Wingender
CEO
Wolfenbüttel, Am Exer 10b
GmbH http://www.genexplain.com

2. We aim to provide a comprehensive platform of
bioinformatics, systems biological and cheminformatics
tools for a
personalized medicine and pharmacogenomics

3. Some facts about geneXplain:
 Founded in April 2010, starting active business July 2010
 International (German-Russian) shareholder structure
 Managing directors: E. Wingender (CEO), A. Kel (CSO)
 Product portfolio in bioinformatics, systems biology,
cheminformatics
 Development close to science and research
 Participation in international and national research consortia
- SYSCOL (EU FP7)
- GERONTOSHIELDS (BMBF)

4. proteins
compounds networks
genes

5. Some facts about geneXplain:
 Founded in April 2010, starting active business July 2010
 International (German-Russian) shareholder structure
 Managing directors: E. Wingender (CEO), A. Kel (CSO)
 Product portfolio in bioinformatics, systems biology,
cheminformatics
 Close to science and research
 Participation in international and national research
consortia
- SYSCOL (EU FP7)
- GERONTOSHIELDS (BMBF)
- TEMPUS (EU)

6. The idea:
 Providing a platform of methods for
 Biomedical research
 Focus: drug development
 Complete pipeline from high-throughput data to a lead structure
 High-throughput data:
 Genomics
 Transkriptomics
 Proteomics
 Public private partnership

7. GeneXplainTM Platform: A Workflow for Drug Discovery
The geneXplain platformTM is a new product integrating bio- and cheminformatics tools for pharmacogenomics.
It provides a drug discovery workflow that guides from the statistical analysis of biological high-throughput
data to a panel of potential lead compounds for further validation.
Statistics
Input: High-throughput data
from patients (genomics,
Within the geneXplain platformTM,
transcriptomics, ChIP-seq,
identification of drug target protein
proteomics, etc.)
molecules by bioinformatics and
Output: List of relevant genes or
systems biology methods, is
proteins
complemented by prediction of
Any pre-processed list of biological activities and adverse
genes or proteins from
Bioinformatics effects for chemical compounds,
own experiments, from
Search for regulatory modules in any literature or databases based on multilevel neighborhoods
genomic regions of atoms (MNA) descriptors.
Output: List of transcription factors
potentially responsible for the observed
(co-)regulation of genes
Any list of transcription factors;
any list of genes or proteins from
own experiments, from literature
Systems Biology
The workflow or databases to be mapped on
Topological analysis of the networks known pathways
The incorporated statistical upstream of transcription factors,
analyses help to identify relevant simulation of the network behavior,
genes or proteins in the raw patient stratification
data, e.g. those that are Hypotheses about
gene regulators Output: List of potential master regulators
differentially expressed.
essential for the
The Bioinformatics block allows
studied process
to reveal potential regulation of
genes by transcription factors or
miRNAs.
Systems biology approaches
analyze networks of molecular Cheminformatics
events and suggest promising Prediction of biological activities of the
Hypotheses compounds, selection of compounds with
drug target molecules and their about target
mechanisms of action. required effects and without adverse or
molecules and
The integrated PASS tool enables their role in the toxic effects.
to direct compound screening by studied process Output: List of potential lead structures
pre-selection of chemicals with Hypotheses for for validation
desirable and without adverse or validations and clinical
toxic effects. trials
Systematic generation of
statistically significant
hypotheses

8. Proof of concept: Net2Drug consortium
EU FP6, Coordinator: A. Kel
Transcriptomics breast cancer cell line
Statistical evaluation
Integrated bioinformatic analysis
(promoter & pathway analysis)
Systems biological simulation
Cheminformatic identification of candidate drugs

9. Proof of concept: Net2Drug consortium
EU FP6, Coordinator: A. Kel
Transcriptomics breast cancer cell line
Results: Statistical evaluation
Out of 24 million compounds, 16 substances turned out
to be feasibleIntegrated bioinformatic analysis
for experimental testing.
(promoter & pathway analysis)
For 2 compounds, highly specific activities were found.
Systems biological simulation
Cheminformatic identification of candidate drugs

10. GeneXplainTM Platform: A Workflow for Drug Discovery
The geneXplain platformTM is a new product integrating bio- and cheminformatics tools for pharmacogenomics.
It provides a drug discovery workflow that guides from the statistical analysis of biological high-throughput
data to a panel of potential lead compounds for further validation.
Statistics
Input: High-throughput data
from patients (genomics,
Within the geneXplain platformTM,
transcriptomics, ChIP-seq,
identification of drug target protein
proteomics, etc.)
molecules by bioinformatics and
Output: List of relevant genes or
systems biology methods, is
proteins
complemented by prediction of
Any pre-processed list of biological activities and adverse
genes or proteins from
Bioinformatics effects for chemical compounds,
own experiments, from
Search for regulatory modules in any literature or databases based on multilevel neighborhoods
genomic regions of atoms (MNA) descriptors.
Output: List of transcription factors
potentially responsible for the observed
(co-)regulation of genes
Any list of transcription factors;
any list of genes or proteins from
own experiments, from literature
Systems Biology
The workflow or databases to be mapped on
Topological analysis of the networks known pathways
The incorporated statistical upstream of transcription factors,
analyses help to identify relevant simulation of the network behavior,
genes or proteins in the raw patient stratification
data, e.g. those that are Hypotheses about
gene regulators Output: List of potential master regulators
differentially expressed.
essential for the
The Bioinformatics block allows
studied process
to reveal potential regulation of
genes by transcription factors or
miRNAs.
Systems biology approaches
analyze networks of molecular Cheminformatics
events and suggest promising Prediction of biological activities of the
Hypotheses compounds, selection of compounds with
drug target molecules and their about target
mechanisms of action. required effects and without adverse or
molecules and
The integrated PASS tool enables their role in the toxic effects.
to direct compound screening by studied process Output: List of potential lead structures
pre-selection of chemicals with Hypotheses for for validation
desirable and without adverse or validations and clinical
toxic effects. trials
Systematic generation of
statistically significant
hypotheses

11. The cheminformatics portfolio:
 PASS
predicts biological activities of chemical compounds from their structural formulae; assigns
probability values to each activity and identifies those parts of the molecule that are responsible
for this activitiy
 PharmaExpert
mines large amounts of predictions generated by PASS to filter out those compounds that
optimaly fit user-defined requirements
 GUSAR
generates quantitative structure-activity relationship (QSAR) models

12. How to get there:
GeneXplainTM Platform: A Workflow for Drug Discovery
The geneXplain platformTM is a new product integrating bio- and cheminformatics tools for pharmacogenomics.
It provides a drug discovery workflow that guides from the statistical analysis of biological high-throughput
data to a panel of potential lead compounds for further validation.
Statistics
Input: High-throughput data
from patients (genomics,
Within the geneXplain platformTM,
transcriptomics, ChIP-seq,
identification of drug target protein
proteomics, etc.)
molecules by bioinformatics and
Output: List of relevant genes or
systems biology methods, is
proteins
complemented by prediction of
Any pre-processed list of biological activities and adverse
genes or proteins from
Bioinformatics effects for chemical compounds,
own experiments, from
Search for regulatory modules in any literature or databases based on multilevel neighborhoods
genomic regions of atoms (MNA) descriptors.
Output: List of transcription factors
potentially responsible for the observed
(co-)regulation of genes
Any list of transcription factors;
any list of genes or proteins from
own experiments, from literature
Systems Biology
The workflow or databases to be mapped on
Topological analysis of the networks known pathways
The incorporated statistical upstream of transcription factors,
analyses help to identify relevant simulation of the network behavior,
genes or proteins in the raw patient stratification
data, e.g. those that are Hypotheses about
gene regulators Output: List of potential master regulators
differentially expressed.
essential for the
The Bioinformatics block allows
studied process
to reveal potential regulation of
genes by transcription factors or
miRNAs.
Systems biology approaches
analyze networks of molecular Cheminformatics
events and suggest promising Prediction of biological activities of the
Hypotheses compounds, selection of compounds with
drug target molecules and their about target
mechanisms of action. required effects and without adverse or
molecules and
The integrated PASS tool enables their role in the toxic effects.
to direct compound screening by studied process Output: List of potential lead structures
pre-selection of chemicals with Hypotheses for for validation
desirable and without adverse or validations and clinical
toxic effects. trials
Systematic generation of
statistically significant
hypotheses

13. The way:
The geneXplain platform
 Integrated collection of bioinformatic and systems
biological program modules („Bricks“)
 Based on proven BioUML technology
 Statistical analysis of high-throughput data
 Integrated bioinformatic promoter- and network analysis
 Systems biological simulation
 Unified look-and-feel
 Workflow management system
 Pre-defined standard workflows
 Easy integration of own tools and scripts

14. Upstream analysis of causes
Key node

15. The way:
The geneXplain platform
 Integrated collection of bioinformatic and systems biological
program modules („Bricks“)
 Based on proven BioUML technology
 Statistical analysis of high-throughput data
 Integrated bioinformatic promoter- and network analysis
 Systems biological simulation
 Unified look-and-feel
 Workflow management system
 Pre-defined standard workflows
 Easy integration of own tools and scripts

16. The geneXplain platform

17. The geneXplain platform

18. The geneXplain platform

19. The geneXplain platform
Public Private Partnership
 Clash of cultures:
 Cheminformatics: commercial approaches accepted
 Bioinformatics: public domain prevalent (Internet culture)
 Advantages of public-domain services:
 Latest state of the art
 Visibility („marketing“ through publications, conference talks, etc.)
 High acceptance
 Disadvantages of public-domain services:
 No unified look-and-feel
 Low user-friendliness
 Poor support
 Uncertainty on side of users without expertise
 Unsure long-term perspective

20. The geneXplain platform
Public Private Partnership
 The disadvantages of the public domain are advantages of a
commercial offer
 Optimal: combination of free and commercial tools
 Business model:
 Platform with integrated free and proprietary offerings
 Payable access
 Payable support

21. The geneXplain platform
Public Private Partnership
 Advantages for the user
 Standardized interface
 Integrated workflows
 Default parametrizations byexperts
 Selection of free modules by experts in the field
 Selection of proprietary, uszually low-price modules by the user
 Full cost-control by the user

22. www.genexplain.com
Contact:
Edgar Wingender edgar.wingender@genexplain.com