The Tenth Annual BioPathways Meeting
Jointly organized with EMERGENCE by
Vítor Martins dos Santos
Vincent Danos
Joanne Luciano
Vincent Schachter
Aviv Regev
Eric Neumann
June 27-28, 2009
As a Special Interest Group Meeting within the ISMB / ECCB 2009
Stockholm, Sweden
The 10th BioPathways meeting is organized by the BioPathways Consortium, an open forum aimed at fostering computational approaches to the understanding of biological networks in biotechnology and biomedicine. Our special focus this year will be on computational methods for Synthetic Biology and will be organised in collaboration with EMERGENCE, an EU-funded consortium aiming at fostering and consolidating the field of Synthetic Biology in Europe.
ftp://ftp.cordis.europa.eu/pub/nest/docs/5-nest-synthetic-080507.pdf
Previous BioPathways meetings have focused on a variety of themes, such as computational reconstruction of molecular networks, pathway evolution, integration of models and experiments, models and ontologies for pathways, modeling of interactions and regulation on a systems scale and network approaches for discovering disease associations, among others.
This year’s meeting will remain within the broad spectrum of computational and semantic approaches for molecular network analysis and systems modeling, featuring presentations also on pathway analysis tools, metabolism, signaling or regulation, selected from the most notable recently published work or from yet unpublished work. In addition, one session each will be dedicated to the two specific thrusts, namely:
a) Design Tools for Synthetic Biology,
b) Computational Pathway Methods for Translational Medicine.
Short descriptions of these topics are as follows:
a) Synthetic Biology addresses the design and fabrication of biological components and systems that do not exist in the natural world as well as the re-design and fabrication of already existing biological systems. In the long run, it would be envisioned to build up cellular components and even cells from scratch to create living devices and use them either as molecular-scale factories, to detect chemical weapons, clean up pollutants, make simple computations, diagnose disease, deliver vaccines, produce energy from water or sunlight, or to create new, hybrid materials. This vision has, potentially, a tremendous scientific, technological and economical impact. Whereas many of the computational methods developed in order to model and analyze the structure and dynamics of natural systems are relevant to the modeling of synthetic living components and systems, there is also an acute need for new computational design methods in order to support the rational design goals and the abstraction/modularity/assembly approach of synthetic biology. The planned session and corresponding round-table discussion aim at contributing to the identification of needs and possibilities.
b) Translational medicine aims broadly at the rapid transformation of laboratory findings into clinically focused applications – ‘from bench to bedside and back’. There has long been a consensus that there is a pressing need to bridge the gap between basic and clinical sciences, to ensure that basic research discoveries of potential relevance to patient care are effectively applied. This is a formidable challenge to implement and some of the key problems stem from the lack of appropriate frameworks and models that link clinically relevant information (in particular that related to multi-scale pathways and networks) to the knowledge obtained across multiple disciplines, experimental platforms and biological systems. The session planned and the corresponding round-table discussion aim at contributing to the identification of major hurdles and possibilities on the development and implementation of computational pathway methods for Translational Medicine. Special emphasis will be put on speakers from industry to illustrate and discuss “translational” issues from the industrial/medical practice.
The meeting will consist of two full days with a balance of invited talks, short oral presentations, and critical, topic-focused panel discussions that will conclude each day. The program and abstracts of the talks are presented below.
We look forward to seeing you at the BioPathways 2009
The organizers: Vítor Martins dos Santos, Helmholtz Centre for Infection Research, Germany (chair), Vincent Schachter, Genoscope, France, Vincent Danos, University Edinburgh, Joanne Luciano, Harvard Medical School and Predictive Medicine Inc., Eric Neumann, Clinical Semantics Group and Aviv Regev, Broad Institute and MIT.
Contact:
Vítor Martins dos Santos, Helmholtz Centre for Infection Research, Germany (chair) vds@helmholtz-hzi.de
10th BioPathways Meeting Program PDF Version Download
Day I – June 27th
Stockholm International Fairs, Room C8
|
7:30 – 8:40 |
Registration |
|
|
8:40-8:55 |
VÍtor Martins dos Santos, Helmholtz Centre for Infection Research (HZI), Braunschweig, Germany |
Opening remarks |
|
Session 1: Pathways Analysis: Databases & Software Tools |
||
|
8:55-9:35 |
Ron Pinter, Israel Institute of Technology, Haifa, Israel |
Conditional pathway integration |
|
9:35-10:15 |
Nicolas Le Novére, EMBL, Cambridge, UK |
The Systems Biology Graphical Notation |
|
10:15-10:45 |
Coffee Break |
|
|
10:45-11:30 |
Satoru Miyano, University of Tokyo, Japan |
Cell illustrator online: A computational platform for systems biology |
|
11:30-12:00 |
Daniele Merico, University of Toronto, Canada |
Enrichment Map: A novel visualization method for gene-set enrichment analysis results |
|
12:00-12:30 |
Rainer König, University of Heidelberg, Germany |
Analyzing enrichments of differentially regulated metabolic enzymes using pathway topology |
|
12:30-13:30 |
Lunch |
|
|
Session 2: Network Reconstruction & Analysis |
||
|
13:30-14:00 |
Luis de Figueiredo, Friedrich-Schiller University Jena, Germany |
Computation of elementary flux modes in systems biology |
|
14:00-14:30 |
Peter Gennemark, University of Göteborg, Sweden |
Formal representation of the high osmolarity glycerol pathway in yeast |
|
14:30-15:00 |
Russel Harmer, CNRS, Paris, France & Harvard University, USA |
Rule-based modelling: Model perturbation and resolution |
|
15:00-15:30 |
Tijana Milenković, University of California at Irvine, CA, USA |
Topological network alignment uncovers biological function and phylogeny |
|
15:30-16:00 |
Coffee Break |
|
|
16:00-16:30 |
Ian Donaldson, University of Oslo, Norway & University of Toronto, Canada |
Providing feedback to protein interaction databases |
|
16:30-17:00 |
Anna Bauer-Mehren, Pompeu Fabra University, Barcelona, Spain |
From SNPs to pathways: Integration of functional effect of sequence variations on models of cell signalling pathways |
|
17:00-17:30 |
Inkyung Jung, KAIST, Daejeon, |
PostNet: Inferring post-translational modification network for multi-target drug development |
|
17:30-18:00 |
Teresa Przytycka, NCBI, Rockville, MD, USA |
Application of combinatorial optimization to predict and study domain-domain interactions |
|
General Discussion |
||
|
18:00-18:30 |
Network analysis, Databases & Tools |
|
Day II – June 28th
Stockholm International Fairs, Room C8
|
Session 3 : Computational Methods and Infrastructure for Synthetic Biology |
|
||
|
8:30-9:00 |
Vítor Martins dos Santos, Helmholtz Centre for Infection Research, Braunschweig, Germany |
EMERGENCE: A foundation for synthetic biology in Europe |
|
|
9:00-9:40 |
Alfonso Valencia, CNIO, Madrid, Spain |
Platforms connecting synthetic and molecular biology resources |
|
|
9:40-10:15 |
Mario Marchisio, ETH Zürich, Switzerland |
Computational design of synthetic gene circuits with composable parts.
|
|
|
10:15-10:45 |
Coffee Break |
|
|
|
10:45-11:25 |
Maria Suárez Diez, Ecole Polytechnique, CNRS, Paris, France
|
Computational design in synthetic biology |
|
|
11:25-12:00 |
Eric Alm, MIT, Cambridge, MA, USA |
New algorithms and approaches for the emerging field of ecological genomics |
|
|
Session 4: Computational Methods in Translational Medicine Chair: Joanne Luciano, Predictive Medicine, Inc. |
|
||
|
12:00-12:30 |
Thomas Pfeiffer, Harvard University, Cambridge, MA, USA |
Estimating publication bias for multiple datasets with similar bias |
|
|
12:30-13:30 |
Lunch |
|
|
|
13:30-14:15 |
Tomer Shlomi, Israel Institute of Technology, Tel-Aviv, Israel |
Predicting metabolic biomarkers of human inborn errors of metabolism |
|
|
14:15-14:50 |
Joanne Luciano, Predictive Medicine, Inc. MA, USA |
EPOS: Enhanced Pathway Optimal Systems |
|
|
14:50-15:30 |
Sven Nelander, University of Gothenburg, Sweden |
Models from patients: in vivo reverse engineering of glioblastoma tumors |
|
|
15:30-16:00 |
Coffee Break |
|
|
|
16:00-16:30 |
Michael Kuhn, EMBL, Heidelberg, Germany |
From protein–chemical interaction networks to human phenotypes |
|
|
16:30-17:00 |
Erik Sonnhammer, Stockholm University, Sweden |
Using FunCoup to find novel disease genes |
|
|
17:00-17:30 |
Eric Neumman, Clinical Semantics Group, Lexington, MA, USA |
An ontology and knowledge base for merging pathway information with genomic and text-mined data |
|
|
Round Table Discussion |
|
||
|
17:30-18:00 |
Computational Methods in Translational Medicine |
|
|
|
End of meeting |
|
||
Conditional pathway integration
Alexandra Skolozub1, Ofer Sarig2 and Ron Y. Pinter1,2
1Department of Computer Science, Technion – Israel Institute of Technology, Haifa, Israel
2Rappaport Institute for Research in the Medical Sciences, the School of Medicine,
Technion – Israel Institute of Technology, Haifa, Israel
Contact: pinter@cs.technion.ac.il
Abstract
Motivation: Many biological pathways that describe complex cellular processes are available in public and commercial databases as well as in the literature. However, each item focuses on a particular cellular function. Moreover, pathways differ in the way they are described in different sources, emphasizing complementary aspects of the biological system under study. Considering related pathways in a unified framework is essential for understanding their behavior and for elucidating and refining open issues involving such systems.
Results: We developed a conditional pathway algebra, in which pathways are enriched with both new node types as well as additional edge types providing significantly more expressive power for the description of existing biological phenomena. During conditional pathway integration, some interactions are made dependent upon a specific predicate (the presence/absence of protein, extracellular factors, etc.). Moreover, such integration enables us to distinguish between different data sources and points out problematic interactions in the given pathways. We provide a formal definition of the algebra and prove some of properties of ist operations, such as closure, commutativity, and the lack of associativity. Some of these operations are essential when applied to several pathways to form an entire (sub)system.
Our algebra is embodied in the Pathway Integration Environment (PIE) as a plug-in for Cytoscape.
To demonstrate the utility and effectiveness of our method, we have applied it to three well characterized yeast signaling pathways: (i) Pheromone response, (ii) Filamentous growth, and (iii) High osmolarity glycerol pathways. Most of our computational observations are confirmed in the literature.
The Systems Biology Graphical Notation
Nicolas Le Novére, on the behalf of the SBGN community
Abstract
Circuit diagrams and UML diagrams are just a few examples of standard visual languages that help accelerate work by promoting regularity, removing ambiguity, and enabling software tool support for communication of complex information. Ironically, despite having one of the highest ratios of graphical to textual information, biology still lacks standard graphical notations. The recent deluge of biological knowledge makes addressing this deficit a pressing concern. Toward this goal, we present the Systems Biology Graphical Notation (SBGN), a visual language developed by a community of biochemists, modelers and computer scientists. SBGN consists of three complementary languages: Process Diagram, Entity Relationship Diagram, and Activity Flow Diagram. Together they enable scientists to represent networks of biochemical interactions in a standard, unambiguous way. We believe that SBGN will foster efficient and accurate representation, visualization, storage, exchange, and reuse of information on all kinds of biological knowledge, from gene regulation, to metabolism, to cellular signaling.
Cell Illustrator Online: A computational platform for systems biology
Masao Nagasaki, Ayumu Saito, Satoru Miyano
Human Genome Center, Institute of Medical Science, The University of Tokyo,
4-6-1 Shirokanedai, Minatoku, Tokyo 108-8639, Japan
Abstract
We developed an Cell System Ontology (CSO) (http://www.csml.org/) and Cell System Markup Language (CSML) and for biological pathways for visualization, modeling and simulation. Based on CSO and CSML, we developed a modeling and simulation tool Cell Illustrator Online (CIO, https://cionline.hgc.jp/) that enables wet lab biologists to draw, model, elucidate and simulate complex biological processes and systems such as metabolic pathways, signal transduction cascades, gene regulatory pathways and dynamic interactions of various biological entities such as genomic DNA, mRNA and proteins. Its architecture employs Hybrid Functional Petri Net with extension (HFPNe) which is defined by enhancing some functions to hybrid Petri net so that various aspects in biological pathways can be intuitively modeled. CIO comes preloaded with TRANSPATH® pathways and chains, providing immediate access to signal transduction and metabolic pathway representations derived from the scientific literature. The integration of TRANSPATH reactions provides direct access to thousands of experimentally demonstrated binding and regulatory relationships – providing a unique set of building blocks for drawing custom networks and pathways. Furthermore, we have developed a method for automatic parameter estimation for HFPNe models by using a technology called data assimilation which "blends" simulation models and observational data "rationally". This data assimilation method is more suited for high performance computing systems and we plan to enroll this function into CI in the near future for high performance computing environment.
Enrichment Map: A novel visualization method for gene-set enrichment analysis results
Daniele Merico, Ruth Isserlin, Oliver Stueker, Andrew Emili, Gary Bader
Terrence Donnelly Centre for Cellular and Biomolecular Research, University of Toronto, Ontario, Canada
Abstract
Gene-set enrichment analysis is a very popular technique enabling the functional characterization of large gene lists, such as high-throughput gene expression results. The large number and high redundancy of gene-sets found in the typical enrichment analysis can obscure the biological functions enriched in the gene list. To overcome this problem, we have developed a novel visualization method, Enrichment Map, which displays gene-sets as networks, where edges between gene-sets represent mutual overlap. Consequently, inter-related gene-sets form network clusters, enabling the user to identify the major functional groups and more easily interpret the enrichment results. A color gradient is used to map enrichment significance.
As an application example of use, we generated the enrichment map for a microarray experiment comparing estradiol-treated and untreated MCF7 breast cancer cells, at two different time points (12 and 24 hours of culture). Inner and outer node colors represent enrichment results at 12 and 24 hours, respectively. Figure 1 displays an excerpt from the enrichment map. The major cluster (red) is formed by several gene-sets related to cell proliferation, all enriched in estradiol-treated cells, in accordance with the known effect of estrogens as stimulators of cell proliferation. The smaller cluster (blue) is related to junctions and it is enriched in untreated cells. Enrichment status is usually consistent between the two time-points; however, for some groups of related gene-sets, induction is maximal at 24 hours (e.g. ubiquitin-dependent protein degradation during cell cycle).
Enrichment Map is implemented as a freely available friendly Cytoscape plug-in.
Figure 1. Enrichment map for estradiol-treated and untreated MCF7 cells.
Analyzing enrichments of differentially regulated metabolic enzymes
using pathway topology
Gunnar Schramm1, Stefan Wiesberg3, Nicolle Diessl1, Anna-Lena Kranz1, Vitalia Sagulenko5, Marcus Oswald3, Gerhard Reinelt3, Frank Westermann5, Roland Eils1 and Rainer König1
1 Department of Bioinformatics and Functional Genomics, Institute of Pharmacy and Molecular Biotechnology,
and Bioquant, University of Heidelberg, Heidelberg, Germany
3 Institute of Computer Science, University of Heidelberg, Heidelberg, Germany
5 Department of Tumor Genetics, German Cancer Research Center (DKFZ),
Heidelberg, Germany
Contact: r.koenig@dkfz.de
Abstract
Gene expression profiling by microarrays or transcript sequencing enables observing the pathogenic function of tumors on a mesoscopic level. We developed a novel algorithm for investigating neuroblastoma tumors which clinically exhibit a very heterogeneous course ranging from rapid growth with fatal outcome to spontaneous regression. In contrast to common enrichment tests, we took network topology into account by applying adjusted wavelet transforms on an elaborated representation of curated pathway maps from KEGG. Our method showed an impressively higher sensitivity when compared to normal enrichment tests. We detected regulatory oncogenetic shifts in the metabolic network of the aggressive form of the tumors. They showed regulatory shifts for purine and pyrimidine biosynthesis as well as folate-mediated metabolism of the one-carbon pool in respect to increased nucleotide production. Interestingly, we also detected a significant expression change in glutamate metabolism. The regulation of this pathway showed a cellular switch of the aggressive tumors for which we provided experimental validation with neuroblastoma cell lines, being the first steps towards new possible drug therapy. In principle, the pattern recognition method we developed can be used for every cellular network and corresponding high throughput datasets. It is superior to normal enrichment tests as it offers a much higher sensitivity for detecting functionally related regulation patterns. We made it publically available as the R-package PathWave (at www.ichip.de).
Computation of elementary flux modes in systems biology
Luis de Figueiredo 1,2, LF, Podhorski 3, A, Rubio 3, A, Kaleta 1, C, Behre 1, J, Beasley 4, JE, Schuster 1, S, and Planes 3, FJ
1 Friedrich-Schiller-University Jena, Jena, Germany
2 PhD Program in Computational Biology, Instituto Gulbenkian de Ciência, Oeiras, Portugal
3 CEIT and TECNUN (University of Navarra), San Sebastián, Spain
4 Brunel University, Uxbridge, United Kingdom
Abstract
With the emergence of Systems Biology, the need for methods that are able to compute pathways in large-scale networks has become increasingly urgent. An elementary flux mode (EFM) is a mathematical representation of a metabolic pathway. Until quite recently, the computation of EFMs in genome-scale models was not possible due to limitations in the existing algorithms. These require the computation of the full set of EFMs, for which the number increases exponentially with network size. We will present some of the recent developments in the computation of EFMs. The improvements in EFM computation increase the applicability of this mathematical framework, reaching a cellular scale. They bring also more flexibility to modelling by focusing on a specific solution of interest. Therefore, EFM analysis is an excellent example of a fruitful application of mathematics to biology, which is essential in the Systems Biology era and, in the other direction, it has even posed new challenges to developments in mathematics, especially in convex analysis.
Formal representation of the high osmolarity glycerol pathway in yeast
Peter Gennemark
Mathematical sciences, University of Göteborg, Sweden
Department of mathematics, Uppsala University, Sweden
Abstract
The high osmolarity glycerol (HOG) signalling system in yeast belongs to the class of Mitogen Activated Protein Kinase (MAPK) pathways that are found in all eukaryotic organisms. It includes at least three scaffold proteins that form complexes, and involves reactions that are strictly dependent on the set of species bound to a certain complex. The scaffold proteins lead to a combinatorial increase in the number of possible states. To date, representations of the HOG pathway have used simplifying
assumptions to avoid this combinatorial problem. Such assumptions are hard to make and may obscure or remove essential properties of the system.
I will discuss a detailed generic formal representation of the HOG system without such assumptions, showing the molecular interactions known from the literature. The model takes complexes into account, and summarises existing knowledge in an unambiguous and detailed representation. It can thus be used to anchor discussions about the HOG system.
In the commonly used Systems Biology Markup Language (SBML), such a model would need to explicitly enumerate all state variables. The Kappa modelling language which we use supports representation of complexes without such enumeration. To conclude, we compare Kappa with a few other modelling languages and software tools that could also be used to represent and model the HOG system.
Rule-based modelling: Model perturbation and resolution
Russel Harmer
CNRS, Paris
Abstract
We will present briefly the framework of rule-based modelling and explain how it solves the combinatorial explosion caused by complex formation and the post-translational modifications typical of biochemistry. We will then generalize this approach to enable a uniform treatment of agent variants. This allows us to avoid a second combinatorial explosion and provides a general framework for building large-scale models and describing model perturbations, e.g. by mutations and/or ligand or drug interventions. In particular, it allows us to test consensus pathways by perturbing them with agents representing the engineered mutations used to infer the pathways in the first place.
Topological network alignment uncovers biological function and phylogeny
Oleksii Kuchaiev, Tijana Milenković, Vesna Memišević, Wayne Hayes, and Nataša Pržulj
Department of Computer Science, University of California, Irvine, CA, USA
Contact: natasha@ics.uci.edu
Abstract
Sequence comparison and alignment has had an enormous impact on our understanding of evolution, biology, and disease. Comparison and alignment of biological networks will likely have a similar impact. Comparing networks of different species has already provided some insight into conservation of proteins, protein-protein interactions, and protein complexes through evolution. However, existing network alignments use information external to the networks, such as protein sequence, because no good algorithm for purely topological alignment has yet been devised. The mass of currently available biological network data will only increase. Since network topology provides a new and independent source of biological information, we believe that high-quality, purely topological alignments can yield new and essential insights into function, evolution, and disease.
Our new algorithm GRAAL (GRAph ALigner) is based solely on network topology, and produces biologically relevant and by far the most complete topological alignments of biological networks to date. We demonstrate that both species phylogeny and detailed biological function of individual proteins can be extracted from a solely topological comparison of networks. Our topological alignment of the metabolic networks of protists and fungi produces phylogenetic trees similar to those based on sequence. Our alignment of the protein-protein interaction networks of two very different species—yeast and human—indicates that even distant species share a surprising amount of network topology with each other, suggesting broad similarities in internal cellular wiring across all life on Earth. GRAAL can easily be applied to other network types, such as gene regulatory, signal transduction, or protein structure networks.
Providing feedback to protein interaction databases
Sabry Razick1, Brian Turner2, Emerson Cho2, Kyle Morrison2, Shoshana Wodak2,3,4 and Ian Donaldson1
1 The Biotechnology Centre of Oslo, University of Oslo, Norway
2 Molecular structure and function program, Hospital for Sick Children, Toronto, Canada
3 Department of Molecular Genetics, University of Toronto, Canada
4 Department of Biochemistry, University of Toronto, Canada
Abstract
Protein interaction data hold incredible potential for biomedical research. These data are collected and archived by a growing number of groups around the world.
As such, it is important that these databases have the means to effectively exchange and compare data and that they are curating and representing data using similar standards to ensure data accessibility and effectiveness.
The iRefIndex project has three long term objectives (http://irefindex.uio.no):
First, facilitate exchange of interaction data between interaction databases.
The iRefIndex paper describes a method for assigning unique and global identifiers to protein interactors, interactions and complexes. This method may be used by anyone to facilitate exchange and consolidation of data.
Second, consolidate interaction data from multiple sources.
The method has been used to index interaction records from ten databases so far. The resulting iRefIndex resource may be used to search for the existence of interaction data for any protein regardless of the original resource.
Third, provide feedback to source interaction databases.
During the process of data consolidation, iRefIndex uses a rigorous method to keep track of potential problems with source records. These data are provided as feedback files to source interaction databases and via a web interface (http://wodaklab.org/iRefWeb) that compares interactions curated by multiple databases on the same publication. This process will help to harmonize data representation and improve the overall quality of interaction records and the ability to compare and exchange records.
From SNPs to pathways: Integration of functional effect of sequence variations on models of cell signalling pathways
Anna Bauer-Mehren, Laura I Furlong, Ferran Sanz
Research Unit on Biomedical Informatics (GRIB), IMIM-Hospital del Mar, Universitat Pompeu Fabra
Barcelona Biomedical Research Park (PRBB), Barcelona, Spain
Contact: ABM: abauer-mehren@imim.es; LIF: lfurlong@imim.es; FS: fsanz@imim.es
Abstract
Single nucleotide polymorphisms (SNPs) are the most frequent type of sequence variation between individuals, and represent a promising tool for finding genetic determinants of complex diseases and understanding the differences in drug response. In this regard, it is of particular interest to study the effect of non-synonymous SNPs in the context of biological networks such as cell signalling pathways. UniProt provides curated information on the functional and phenotypic effects of natural variation, including SNPs, as well as on mutations on protein sequences. However, no strategy has been developed to integrate this information with biological networks, with the ultimate goal of studying the impact of the functional effect of SNPs in the structure and dynamics of biological networks.
We identified the different challenges posed by the integration of the phenotypic effect of natural sequence variants and mutations with biological pathways. Moreover, we developed a strategy for the combination of data extracted from public resources, such as UniProt, NCBI dbSNP, Reactome and BioModels. We generated attribute files containing phenotypic and genotypic annotations to the nodes of biological networks, which can be imported into network visualization tools such as Cytoscape.
These resources allow the mapping and visualization of mutations and natural variations of human proteins and their phenotypic effect on biological networks. We expect that this approach will help in the study of the functional impact of disease associated SNPs in the behavior of cell signaling pathways, which ultimately will lead to a better understanding of the mechanisms underlying complex diseases.
PostNet: Inferring post-translational modification network for
multi-target drug development
Inkyung Jung1 and Dongsup Kim1,2*
1Department of Bio and Brain Engineering and
2KAIST Institute for BioCentury, KAIST, Daejeon 305-701, S. Korea
Abstract
A strategy for inhibiting multiple kinases to develop multi-target drugs has become a promising tool as a new type of therapeutics. Protein kinases are critical component of cellular signal transduction cascades and directly involved in many diseases. Several drugs such as Gleeve for gastrointestinal stromal tumor (GIST) and Iressa for non-small cell lung cancer (NSCLC) are examples to support the fact that small molecule kinase inhibitors can be effective to cure disease. In this regard, we suggest an efficient way to determine enzyme inhibitors specificity for each disease by inferring pos-translational modification (PTM) network. We integrated the public databases for PTM relationships, binding relationships between small molecules and proteins, and disease relationships with genes. We screened possible specific inhibitors as drug candidates by measuring the relationship between enzymes and diseases via PTM network. From the relationships we noted that considering PTM network is useful for selecting disease specific enzymes for development of multi-target drugs. We compared the final results of chemical-disease relationships with seven FDA approved kinase inhibiting drugs (Gleevec, Lressa, Tarceva, Nexava, Sutent, Sprycel, Tykerb). Most of drugs were highly ranked in a specific disease (the rank of Gleevec for GIST is second and that of Iressa for NSCLC is 55th among 17771 chemicals). Furthermore, we suggested new drug candidates and found a potential new role of FDA approved drugs for the other diseases. Our systematic approach can be applied to identifying an effective set of protein targets for multi-target drugs and drug repositioning.
Application of combinatorial optimization to predict and study domain-domain interactions
Teresa Przytycka
NCBI /NLM / NIH
Rockville, MD, USA
Abstract
Comprehending the cell functionality requires knowledge about the functionality of individual proteins as well as the interactions among them. Proteins typically contain two or more domains, and a protein interaction usually involves binding between specific pairs of domains. Identifying such interacting domain pairs is an important step towards determining the protein-protein interaction network. We demonstrate that evolutionary parsimony principle combined with combinatorial optimization techniques leads to an approach to detecting domain-domain interactions that outperforms other methods to attack the problems.
Joint work with Katia Guimaraes, Elena Zotenko and Raja Jothi.
EMERGENCE: A Foundation for Synthetic Biology in Europe
Vítor Martins dos Santos on behalf of the consortium EMERGENCE
www.emergence.ethz.ch/
Abstract
Synthetic biology has emerged as a very recent but highly promising approach to re-organizing the scientific biological endeavor by integrating central elements of engineering design. By applying the tool box of engineering disciplines such as electrical, mechanical, or chemical engineering and computer sciences, including the vigorous application of modeling techniques and organizing the development of novel biological systems along a hierarchical systems architecture with defined and standardized interfaces, synthetic biology aims at no less than revolutionizing the way we do bioengineering today. If successful, synthetic biology will transform bioengineering into a highly successful and sustainable life science industry.
The objective of the coordination action (CA) EMERGENCE is to provide a communication and working platform for the emerging European synthetic biology community in order to strengthen the organizational and conceptual basis of the synthetic biology as a true engineering discipline in biological engineering.
These issues are addressed in terms of
1. Integration, e.g., providing an organizational forum for the various ongoing activities in the field of synthetic biology (projects in the NEST calls under the synthetic biology initiative).
2. Common concepts and agenda, e.g., providing a common IT-infrastructure to include data sets relevant to synthetic biology as well as tools dedicated to biological design.
3. Standardization, e.g., implementing standards and gene regulations to define the meaning of a number of imprecise terms and concepts.
4. Education, e.g., analyzing the case for a European and world-wide community (‘education focus groups’ to coordinate initiatives as participating in the iGEM competition, establishing a ‘European Master of Synthetic Biology’).
5. Embedding industry, e.g., integrating representatives from industry into the synthetic biology community as the implementation of widely accepted standards will facilitate the development of novel industries.
Platforms connecting synthetic and molecular biology resources
Victor de la Torre and Alfonso Valencia
Spanish National Cancer Research Centre (CNIO)
and
Spanish National Bioinformatics Institute (INB)
Contact: valencia@cnio.es
Abstract
We have developed MADAS as an integral annotation system. MADAS is able to combine the very diverse information sources required for genomic projects, add annotations provided by users under a controlled authoring system, and represent the results using a set of flexible modules. We have used MADAS in a number of projects including the annotation of genomes (for example, Pseudomonas in the context of the P-Sysmo project). More recently, in collaboration with Randy Rettberg (MIT), we have used MADAS as a platform for connecting the information on non-biological parts organized in the MIT "repository of parts" with the corresponding similar biological elements contained in various databases.
Computational design of synthetic gene circuits with composable parts
M.A. Marchisio and J. Stelling
D-BSSE, Basel, ETH Zurich, Switzerland
Abstract
Novel genetic circuits can be engineered using standard parts with well-understood functionalities. However, so far no model based on the simple composition of these parts has become a standard. This is mainly because of the difficulty of defining signal exchanges between biological units unambiguously as it is possible in electrical engineering. Taking inspiration from (and slightly modifying) ideas in the ”MIT Registry of Standard Biological Parts”, we developed a method for the design of genetic circuits with composable parts. Gene expression requires five kinds of signal carriers: RNA polymerases, ribosomes, transcription factors, small RNAs and chemicals (inducers and corepressors). The flux of each of these types of molecules is regarded as a quantifiable biological signal exchanged between parts. Here, each part is modeled independently by the ordinary differential equations (ODE) formalism and integrated into the software ProMoT (Process Modeling Tool- http://www.mpi-magdeburg.mpg.de/projects/promot/). In this way we realized a ”drag and drop” tool, where genetic circuits are built just by placing biological parts on a canvas and by connecting them through ”wires” that enable flow of signal carriers, as it happens in electrical engineering. Besides parts, we considered pools of free signal carriers. They permit to better depict interactions among transcription units and to study the circuit scalability with respect to the number of components. Our simulations of well-known synthetic circuits agree well with published computational and experimental results.
Computational design in synthetic biology
María Suárez Diez1,2, Pablo Tortosa Martínez1, Javier Carrera3, Guillermo Rodrigo3, Alfonso Jaramillo1,2
1. Laboratoire de Biochimie, Ecole Polytechnique, CNRS France
2. SYNTH-BIO group. Epigenomics Project, Genopole UEVE CNRS France
3. Instituto de Biología Molecular y Celular de Plantas, CSIC/Universidad Politécnica de Valencia, Spain
Contact: maria.suarez@epigenomique.genopole.fr
Abstract
The rational design of novel biological networks with prescribed functions is limited to gene circuits of a few genes. Larger networks involve complex interactions with many parameters and their design demands the use of computational methods able to incorporate the design principles of biological circuits. We propose a computational procedure to design circuits composed of predefined genetic parts with a targeted function. The proposed circuits are computationally evolved by either modifying their topology or their kinetic parameters, and by using a fitness function to select for a targeted dynamics. We have applied this automatic approach to design, using transcription networks, digital circuits, oscillators and memory devices. The design of functional biological circuits also requires the design and construction of well characterized biological modules with context-independent targeted behavior and we are often confronted by the task of designing new parts and devices composed of proteins with specified functionalities that are not found in nature. We will discuss the applications of our computational protein design methodology to design proteins with new functionalities. Additionally, we have developed a method for the automatic assembly of models of biological parts that will allow not only the use of arbitrary kinetic models but also the incorporation of experimental data back into the design process. We have also developed an automated method aimed at the de novo design of metabolic pathways using a retrosynthetic algorithm. This tool will allow grafting new bioproduction pathways into a given cellular chassis.
New algorithms and approaches for the emerging field of
ecological genomics
Eric Alm
MIT, Cambridge, MA, USA
Abstract
I will discuss some of the key challenges and opporunities for computational biologists in the emerging field of ecological genomics. I will focus on applications related to microbial ecology, and present several new algorithms and approaches recently developed in my group.
Estimating publication bias for multiple datasets with similar bias
Thomas Pfeiffer
Program for Evolutionary Dynamics, Harvard University;
Contact: pfeiffer@fas.harvard.edu
Abstract
The synthesis of results from published studies in meta-analyses and systematic reviews plays an increasing role in evidence-based medicine. Published results, however, are not necessarily typical results because studies that yield a statistically significant outcome might have an increased chance to get published. The resulting bias may distort the outcome of a meta-analysis and is one of the major obstacles in evidence synthesis. Statistical methods have been developed to detect and correct for publication bias. However, a detailed modeling requires extensive data. Such data can be generated by combining datasets with a similar bias. I present an extension of previous statistical approaches to quantify publication bias from multiple datasets and employ it on data from case-control studies on the genetic basis of Alzheimer’s disease.
Predicting metabolic biomarkers of human inborn errors of metabolism
Tomer Shlomi, Moran Cabili and Eytan Ruppin
CS faculty in the Technion, Israel Institute of Technology, Tel-Aviv University
Abstract
Early diagnosis of inborn errors of metabolism is commonly performed through biofluid metabolomics, which detects specific metabolic biomarkers whose concentration is altered due to genomic mutations. The identification of new biomarkers is of major importance to biomedical research and is usually performed through data mining of metabolomic data. After the recent publication of the genome-scale network model of human metabolism, we present a novel computational approach for systematically predicting metabolic biomarkers in stochiometric metabolic models. Applying the method to predict biomarkers for disruptions of red-blood cell metabolism demonstrates a marked correlation with altered metabolic concentrations inferred through kinetic model simulations. Applying the method to the genome-scale human model reveals a set of 233 metabolites whose concentration is predicted to be either elevated or reduced as a result of 176 possible dysfunctional enzymes. The method’s predictions are shown to significantly correlate with known disease biomarkers and to predict many novel potential biomarkers. Using this method to prioritize metabolite measurement experiments to identify new biomarkers can provide an order of a 10-fold increase in biomarker detection performance.
EPOS: Enhanced Pathway Optimal Systems - An exploratory
Joanne Luciano
Predictive Medicine, Inc.
Abstract
There is a disconnection in contemporary life science research between modelers, who construct in silico representations of organic phenomena, and experimentalists, who conduct bench or clinical research. The value of a model lies in its ability to explain or make predictions about phenomena it is modeling, and yet it is only data from experiments that can validate or reveal inconsistencies in the model. This project aims to bridge this gap specifically for research in diabetes research. Outcome of this work is intended to facilitate drug target identification, patient population identification and biomarker identification.
In this talk I will present the EPOS approach. The goal of the EPOS project is to explore the creation of a consistent model of insulin resistance in adipose tissue which can be used, for example, for target identification. The approach bases the model on experimental data of various types (microarray, mass spectroscopy) utilizing a number of emerging technologies. These technologies include workflows (Taverna, i2b2 hive), visualization environments (CellDesigner), ontologies (OWL) and reasoners (Pellet). EPOS is intended to provide the basis for continued research in diabetes by creating a metabolic model of insulin resistance that can be constrained by clinical data measurements. This will enable diabetes researchers to use their model to predict experimental results and explain their hypotheses. EPOS is exploratory because it is unclear whether the representational technologies will fully capture the contextual representation of the experimental data clinical researchers provide. And, equally, whether the experimental techniques themselves need refinement, as it is difficult to measure the many relevant phenomena. The outcome, apart from the model, will be a deeper understanding of insulin resistance directly and a greater understanding of where to focus experimental and computational tools. Since much of the current research in the life sciences converges on the study of cellular pathways, we will the validate our methodology by its ability to aid in our understanding of insulin resistance in adipose tissue and chart the course for future explorations using these technologies in other pathway focused research in and beyond the scope of diabetes.
Models from patients: in vivo reverse engineering of glioblastoma tumors
Bodil Nordlander, Teresia Dahl, Erik Johansson, Keiko Funa, Peter Gennemark, Sven Nelander
Sahlgrenska Center for Cardiovascular and Metabolic Research, University of Gothenburg, Sweden
Abstract
We propose a way to construct in vivo transcriptional network models for human cancer tumors. The key idea is to view DNA mutations as system perturbations and mRNA expression changes as the system response.
We used a rich data-set from the Cancer Genome Atlas to construct a model based on data from 186 glioblastoma patients. The model is significantly enriched for oncogenes such as EGFR, PDGFA, PDGFRA, MET, ALK (typically at hub positions) and genes involved in neural differentiation and development (typically at downstream positions). Further, we identify three unexpected hub genes: one G-protein coupled receptor, one redox enzyme and one nuclear protein. We perform cell experiments to confirm transcriptional interactions and the impact of hub gene perturbation on growth rate.
In summary, our work shows that mutational profiles from human cancer tumors can be studied in terms of their regulatory impact, rather than their frequency. This could result in better detection of cancer-driving mutations, and may help identify new drug target candidates.
From protein–chemical interaction networks to human phenotypes
Michael Kuhn, Peer Bork
EMBL Heidelberg, Germany
Abstract
The increasing amount of data available on drugs and their molecular and phenotypic effects is creating exciting opportunities for research. We have combined different sources on protein–chemical interaction data into the network database STITCH (http://stitch.embl.de), which contains interaction information for 68,000 chemicals, including 2,200 drugs. Using this data on a set of side effects for 746 drugs, we developed a side effect similarity measure that predicts whether two drugs share targets. Testing this method on 20 candidates, we found activity in 13 cases. To enable other scientists to do research on the phenotypic effects of drugs, we have created SIDER as a resource on side effects (http://sideeffects.embl.de).
Using FunCoup to find novel disease genes
Erik Sonnhammer and Andrey Alexeyenko
Stockholm Bioinformatics Centre
AlbaNova University Centre, Stockholm University, Sweden
Abstract
I will present FunCoup, an optimised Bayesian framework and a web resource for gene/protein network reconstruction and analysis. The framework was comprehensively tested to optimise the overall confidence and ensure seamless, automated incorporation of new data sets of heterogeneous types. Using over 50 datasets in seven organisms, and extensively transferring information between orthologs, FunCoup predicted global networks in eight eukaryotes. FunCoup predictions were validated on independent cancer mutation data, and were applied to discover candidate members of the Parkinson and Alzheimer pathways.
Because interactomes comprise functional coupling of many types, FunCoup annotates network edges with confidence scores in support of different kinds of interactions physical interaction, protein complex member, metabolic or signalling link. This capability boosted overall accuracy.
The networks, which are the largest interactome reconstructions to date of the respective species, are freely available for download and query at http://FunCoup.sbc.su.se. The site allows detailed graphical and tabular analysis of subnetworks around query genes, as well as comparative analysis of aligned orthologous networks in multiple species.
An ontology and knowledgebase for merging pathway information with genomic and text-mined data
Eric Neumann
Director, Clinical Semantics Group
Lexington, MA, USA
Abstract
We describe an application that combines an ontology supporting therapeutic R&D with a knowledge base instantiated with genomic and text-mined data. As such, this knowledge application is able to effectively incorporate and apply newly generated scientific data to the challenges of disease biology whereby new insights can be rapidly made. This kind of data-driven research in the context of current biomedical knowledge is being used in a pharmaceutical R&D environment. The semantic web interface enables scientists to easily explore the complex relations using faceting browsing. The system also supports connections to over 14 external taxonomies and 8 primary data sources (e.g., NCBI, Reactome, Uniprot). It is our belief that ontological system can serve as a bridge between digital resources and domain users who are scientists, rather than technical experts.
8