BIOTEC - Biotechnology Center TU Dresden

Computational Drug Repositioning with Networks, Structures, Text & Ontologies.  

Developing a drug costs in the order of a billion USD and takes some 10 years. Drug repositioning tries to reduce cost and time by applying drugs that are already on the market to new uses. With the avalanche of publicly available data on compounds, genes, proteins, and diseases and their relationships, there is a tremendous need for computational approaches to organize and analyze this data and finally predict novel uses. In our group we are specifically focusing on developing algorithms and analysis pipelines using networks, protein structures, text-mining and ontologies for this purpose. We currently focus on pancreas cancer as application, where we have a wealth of experimental data from collaborations with the medical faculty.

Networks

We construct drug-target-disease networks by integrating numerous public resources into one coherent framework. We then develop graph algorithms to mine these networks and data to predict novel relationships. Concretely, we exploit incomplete bi-cliques in these networks to predict novel uses for anti-parasitic compounds in heart disease and we apply an algorithm similar to Google’s Pagerank to find biomarkers in pancreas cancer.

Drug Repositioning through Incomplete Bi-cliques in an Integrated Drug-Target-Disease Network.
Simone Daminelli, V Joachim Haupt, Matthias Reimann, and Michael Schroeder.
Integrative Biology, 2012

Prediction of Survival and Response to Adjuvant Therapy in Pancreatic Cancer.
Christof Winter, Janine Roy, Glen Kristiansen, Stephan Kersting, Daniela Aust, Thomas Knösel, Petra Rümmele, Beatrix Jahnke, Vera Hentrich, Felix Rückert, Marco Niedergethmann, Wilko Weichert, Marcus Bahra, Hans J. Schlitt, Utz Settmacher, Helmut Friess, Markus Büchler, Hans-Detlev Saeger, Michael Schroeder, Robert Grützmann, and Christian Pilarsky.
PLoS Computational Biology, 2012

Structures

BVDU is a drug against herpes that is already on the market since the 80s. There are clinical trials to test BVDU as a pancreas cancer drug. In herpes, BVDU targets a viral thymidine kinase; in pancreas cancer the heat shock protein Hsp27. For the former, there is a 3D structure; for the latter we built a model. The viral thymidine kinase and the human heat shock protein have absolutely nothing in common, but a pocket with five residues in the same conformation to bind BVDU. Our algorithm identified this pocket and our prediction was experimentally verified. Based on this knowledge, we aim to predict novel compounds that are even more potent than BVDU is. In general, the quest remains to systematically compare potential binding sites in all protein structures to infer novel uses and side effects of drugs.

RP101 (Brivudine) binds to heat shock protein HSP27 (HSPB1) and enhances survival in animals and pancreatic cancer patients.
Jörg-Christian Heinrich, Anne Tuukkanen, Michael Schroeder, Torsten Fahrig, and Rudolf Fahrig.
Journal of Cancer Research and Clinical Oncology, 137(9):1349--61, 2011

Protein interactions in 3D: From interface evolution to drug repositioning.
Christof Winter, Andreas Henschel, Anne Tuukkanen, and Michael Schroeder.
Journal of Structural Biology, 2012

Text-mining and ontologies.

In the 80s, Swanson put two facts from two papers together to formulate a novel hypothesis: One paper said that Raynaud’s syndrome is caused by increase blood viscosity and another said that fish oil reduces blood viscosity. Taken together, fish oil may be beneficial against Raynaud’s disease. Swanson’s discovery was serendipitous. Can we systematically mine scientific literature and patents to automatically make such predictions? We have developed powerful tools, which extract facts from text and allow drug-target prediction and systematic search of large volumes of text.

Computational Polypharmacology with Text Mining and Ontologies.
Conrad Plake and Michael Schroeder.
Current Pharmaceutical Biotechnology, 12(3):449--57, 2011

Extending ontologies by finding siblings using set expansion techniques.
Götz Fabian, Thomas Wächter, and Michael Schroeder.
Bioinformatics, 2012

For other research in computational biology, please see compbio-dresden.de

For a complete list of papers click here 

• Prediction of drug-gene associations via ontological profile similarity with application to drug repositioning.
  Maria Kissa, George Tsatsaronis, and Michael Schroeder.
  Methods, 2015
• Network information improves cancer outcome prediction.
  Janine Roy, Christof Winter, Zerrin Isik, and Michael Schroeder.
  Briefings in Bioinformatics, 15(4):612--25, 2014
• Drug Promiscuity in PDB: Protein Binding Site Similarity is Key.
  Joachim Haupt, Simone Daminelli, and Michael Schroeder.
  PLoS One, 2013
• Drug Repositioning through Incomplete Bi-cliques in an Integrated Drug-Target-Disease Network.
  Simone Daminelli, V Joachim Haupt, Matthias Reimann, and Michael Schroeder.
  Integrative Biology, 2012

Prof. Michael Schroeder, michael.schroeder(at)biotec.tu-dresden.de, 
Tel: 0049 351 463 400 62
 Assistant: Mandy Glässer, mandy.glaesser(at)biotec.tu-dresden.de, 
Tel: 0049 351 463 400 60
 Address: BIOTEC, TU Dresden, Tatzberg 47-51, 01307 Dresden, Germany Map

We are always looking for excellent people. Please send your CV and why you want to join us. 

GoPubMed

GoPubMed is a novel semantic search engine, which helps answering questions using ontologies. This type of semantic search is also available for web search (GoWeb), gene annotation (GoGene) and search for alternatives to animal experiments (Go3R). 

 

 

 

 

SCOPPI

 SCOPPI is a database of structural protein interactions. 

 

 

 

 

PLIP

The Protein-Ligand Interaction Profiler (PLIP) is a web service and command line tool for fully automated characterization of non-covalent interactions between proteins and ligands in 3D structures.

 

 

PowerGraphs

 PowerGraphs is a software to find modules in graphs. 

 

 

 

DOG4DAG

DOG4DAG is an ontology generation plugin for Protégé and OBO-Edit that allows users to automatically generate ontology terms, definitions and parent-child relationships. 

 

metaPocket 2.0

 

Transinsight is a spin-off founded in late 2005, which markets and develops GoPubMed and related software.

• Transinsight wins German Innovation Prize at Cebit (43 winners out of 2000 submissions) , 3/2011
• Nomination of Dr. Loic Royer for PhD thesis award of German Society of Computing (GI Dissertationspreis). 2.2011
• Christof Winter wins Klaus Tschira prize, Dresden, 10.2010
• GoPubMed wins Deutschen Industriepreis 2010 in category research and development, Hannover. 19.4.2010
• Dr. Andreas Doms wins Georg Helms prize of TU Dresden. 10.2009
• Transinsight wins Red Dot Design Award, Wuppertal, Germany, 8.2009
• Nanometis wins 2nd prize Science4Life businessplan competition, Frankfurt, 7.2009
• Nanometis wins 1st prize FutureSax businessplan competition, Leipzig, 7.2009 
• Nanometis wins Science4Life concept competition, Berlin, 3.2009
• Nomination of Dr. Andreas Henschel for PhD thesis award of German Society of Computing (GI Dissertationspreis). 2.2009
• Lohmann-Medal for diploma thesis of Steffen Jensch. 12.12.2008 
•  Transinsight named among top 5 biotec start ups at European Venture Summit, Düsseldorf, 2.12.2008
• Output-Award of computer science faculty of TU Dresden for GoPubMed. Dresden, 18.4.2008
• 1st prize for presentation of Transinsight at futureSax 'Durchstarten mit Riskikokapital'. Dresden, 17.10.2007 
• BioCreative textmining challenge 2007. Best group in gene name identification task. Madrid, 23.4.2007 
•  Award "Leuchtturmprojekt des High-Tech Gründerfonds" by national ministry for economic affairs and technology for Transinsight.com. Berlin, 16.2.2006
• IBM-prize of computing faculty of TU Dresden for diploma thesis of Andreas Doms, 2005

Current Group

 

Oberwiesentahl Group retreat 2011

 

Alex Mestiashvili, Janine Roy, Joachim Haupt, Jörg Heinrich, Mandy Glässer, Michael Schroeder, Norhan Mahfouz, Simone Daminelli, Sebastian Salentin.

You can find their contact details on the Staff page

Former PhDs and PostDocs

 
• Simone Daminelli, 2010-2016, now at Fios Genomics, Edinburgh
• George Tsatsaronis, 2010-2015, now Elsevier, Amsterdam
• Daniel Eisinger, 2010-2015, now Zencap, Berlin
• Maria Kissa, 2011-2015, now PopSugar, San Francisco, USA
• Eleni Christodoulou, 2011-2015, now National Cancer Center, Singapore
• Zerrin Isik, 2011-2014, now assistant professor at Dokuz Eylül Üniversitesi, Izmir, Turkey
• Thomas Wächter, 2005-2012, now T-Systems, Dresden
• Joscha Köllner, 2009-2011, now KiwiGrid, Dresden 
• Conrad Plake , 2008-20011, now ID Information und Dokumentation im Gesundheitsewsen, Berlin
• Rainer Winnenburg, 2007-2011, now PostDoc, Stanford, USA
• Anne Tuukkanen, 2006-2011, now EMBL, Hamburg
• Dimitra Alexopoulou, 2006-2011, now Biotec, TU Dresden
• Loic Royer, 2005-2011, now PostDoc, MPI-CBG, Dresden
• Frank Dressel, 2005-2011, now Plastic Logic, Dresden
• Christof Winter, 2004-2010, now Groupleader, Munich
• Heiko Dietze, 2005-2010, now Lawrence Berkeley National Laboratory, Berkeley, USA 
• Annalisa Marsico, 2006-2010, now Professor, FU Berlin and  MPI Molecular Genetics, Berlin
• Bill Andreopoulos, 2006-2010, now Lawrence Berkeley National Laboratory, Berkeley, USA
• Andreas Doms, 2004-2009, now SAP Research, Dresden 
• Gihan Dawelbait, 2004-2009, now Masdar, Abu Dhabi
• Andreas Henschel, 2004-2009, now Professor, Masdar, Abu Dhabi
• Dirk Labudde, 2006-2009, now Professor, FH Mittweida 
• Adrian Paschke, 2007/8, now Professor, FU Berlin 
• Bingding Huang, 2004-2007, now Bioinformatics scientist at Computational Oncology Group, German Cancer Research Center, Heidelberg
• Jörg Hakenberg, 2006-2007, now Senior Bioinformatics Scientist, now Illumina, Santa Clara, California, USA
• Wan Kim, 2004-2006, now Professor, Ewha Womans University, Seoul, Korea 
• Nataraj Dongre, 2004/5, now Virginia Tech 

 

 
We organised the national bioinformatics conference GCB2008

 
 2006 - group retreat in saxonia-switzerland 

 
 2007 - summer excursion 
 

 
 2008 - group retreat in Eibenstock 

 
 2011 - Bowling

 
 2013 - Group retreat in Czech Republic

 

Funding by BMBF, BMWi, EU, DFG is kindly acknowledged.

• Fortissimo, EU, 2014-2016
• SigSax, Saxony, 2015-2016
• OpenScienceLink, EU, 2013-2016
• Hybris, DFG, 2012-2015
• GeneCloud, BMWi, 2011-2014
• PONTE, EU, 2010-2013 
• PPI-Marker, EU, 2010-2012 
• GoOn, BMBF, 2009-2011 
• Go3R, BMBF, 2009-2012
• CLSD Computational life science lab, BMBF, 2008-2011
• Zoomable Cell, DFG, 2008-2011
• Nanobrain, BMBF, 2008
• Sealife, EU, 2006-2009
• GoEverywhere, EFRE, 2005-2007
• FoldUnfold, EFRE, 2005-2007
• CODI, EFRE, 2004-2006
• REWERSE, EU, 2004-2008
• BioGrid, EU, 2002-2004
• GeneStream, EU, 2002-2003

We take part in the international Bachelor and Master Courses hold at the TU Dresden and BIOTEC. 
Information about the Bioinformatic classes can be found here.

22. Sainitin Donakonda. Prediction and validation of novel compounds binding the cancer target HSPB1. 2015

21. Maria Kissa. Cassandra: drug gene association prediction via text mining and ontologies. 2015

20. Eleni Christodoulou. Detection of KRAS Synthetic Lethal Partners through Integration of existing RNAi Screens. 2015

19. Simone Daminelli. Drug repositioning with incomplete bi-cliques and structural bioinformatics. 2014

18. Joachim Haupt. Protein binding site similarity as driver for drug repositioning. 2014

17. Janine Roy. From correlation to causality: Outcome prediction in cancer. Summa cum laude. 2014

16. Daniel Eisinger. Automated Patent Categorisation and Guided Patent Search using IPC as inspired by MeSH and PubMed. 2013

15. Rainer Winnenburg. Mutations and their role in protein interactions and diseases: Computational methods for automated retrieval, analysis, and application. 2011

14. Conrad Plake. Gene annotation by automated literature analysis with an application to drug-target interaction prediction. 2010 

13. Loic Royer. Unravelling the structure and assessing the quality of protein networks with power graphs. Summa cum laude. 2010 

12. Thomas Wächter. Semi-automated ontology generation for biocuration and semantic search. Summa cum laude. 2010 

11. Heiko Dietze. GoWeb: Semantic search and browsing for the life sciences. 2010 

10. Dimitra Alexopoulou. Word-sense disambiguation in biomedical ontologies. 2010 

9. Annalisa Marsico. Discovery and characterization of new structure and function motifs in alpha-helical transmembrane proteins with an application to interpret Single Molecule Force Spectroscopy data. Summa cum laude. 2009 

8. Christof Winter. Protein interactions in disease: Using structural protein interaction and regulatory networks to predict disease-related mechanisms. Summa cum laude. 2009 

7. Gihan Dawelbait. From cancer gene expression to protein interaction: Interaction prediction, network reasoning and applications in pancreatic cancer. 2009 

6. Andreas Doms. GoPubMed:ontology-based literature search for life sciences. Summa cum laude. 2008

5. Andreas Henschel. A structural classification of protein-protein interactions for deteection Protein interactions. Summa cum laude. 2008

4. Frank Dressel. Sequenz, Energie, Struktur - Untersuchungen zur Beziehung zwischen Primär- und  Tertiärstruktur in globulären und Membran-Proteinen. 2008 

3. Bingding Huang. Improving protein docking with binding site prediction. 2008 

2. Penny Noy. Enhancing comprehension of Complex Data Visualizations: Framework and Techniques Based on Signature Exploration. 2004. 

1. Jacek Gomoluch. Market Protocols for Resource Allocation in Computational Clusters and Grids. 2003