Extending and Integration of Rule-based Modeling into Virtual Cell Modeling and Simulation Framework
I'm working in close collaboration with Virtual Cell team. I'm developing rule-based applications for modeling in VCell mixing reactions and rules. The latest rule-based enabled version of VCell software is a culmination of the three years of development for all different aspects - user-interface, back-end algorithms and client- and server- based simulations.
- Blinov, Schaff, Vasilescu, Moraru, Bloom & Loew. (2017) Biophys. J.
- Schaff, Vasilescu, Moraru, Loew & Blinov (2017) Bioinformatics
- Moraru, Schaff, Slepchenko, Blinov, Morgan, Lakshminarayana, Gao, Li, & Loew (2008). IET Systems Biology
Algorithms and software for rule-based modeling
Signal transduction networks often exhibit combinatorial complexity: the number of protein complexes and modification states that potentially can be generated during the response to a signal is large, because signaling proteins contain multiple sites of modification and interact with multiple binding partners. An alternative to the conventional approach is a rule-based description, where all potential chemical species and reactions in the model are generated automatically by a computer algorithm from a set of rules. Rule-based modeling method and software BioNetGen can describe a broad range of biological effects.
- Blinov & Moraru (2012) Adv Exp Med Biol.
- Blinov, Yang, Faeder & Hlavacek (2006) LNCS
- Hlavacek, Faeder, Blinov, Posner, Hucka & Fontana (2006) Sci. STKE
- Blinov, Faeder, Yang, Goldstein & Hlavacek (2005) Nat.Biotech.
- Faeder*, Blinov*, Goldstein & Hlavacek (2005) Complexity (* Equal contribution)
- Blinov, Faeder, Goldstein & Hlavacek (2004) Bioinformatics
New Methods for Mathematical Modeling in Biology
Multivalency may lead to the formation of large molecular clusters or polymers even when the individual binding affinities are weak, such as P-granules, mRNA granules, the assembly focal adhesions and the aggregation of receptor signaling platforms. Such interactions lead to the formation of molecular clusters that increase local concentration of biomolecules, potentially triggering signaling events. Because these complexes often have variable composition, we call them pleomorphic ensembles (PEs), to distinguish them from machines, assemblies of strongly and specifically interacting molecules. Large number of molecules in molecular clusters with rapidly changing composition requires new modeling techniques to simulate the dynamics of cluster composition changes.
- Falkenberg, Blinov & Loew (2013) Biophys J. [SuppMat]
- Blinov & Moraru (2012) BMC Biology
- Mayer, Blinov & Loew (2009) J. Biol.
- Blinov, Ruebenacker & Moraru (2008) IET Systems Biology
- Hlavacek, Faeder, Blinov, Perelson & Goldstein (2003) Biotechnol. Bioeng.
Mechanistic Computational Modeling of Signal Transduction
Detailed mechanistic modeling of signal transduction network in a single cell describes activities and interactions among domains of biomolecules (e.g., phosphorylation of specific tyrosine residues, interactions between SH2 domain and phosphotyrosine). I mostly deal with epidermal growth factor receptor signaling, but I was also involved in modeling of signaling by high-affinity immunoreceptors, and looked into various signaling systems, e.g., notch pathway, insulin growth factor receptor signaling, interleukin-1 (IL-1) and toll-like receptor signaling.
- Kesseler, Blinov, Elston, Kauffman & Simpson (2013) J Theor Biol
- Nag, Monine, Goldstein, Faeder & Blinov (2012) "Protein Kinases / Book 2"
- Nag, Monine, Blinov, Goldstein (2010) J.Immunol.
- Blinov, Faeder, Goldstein & Hlavacek (2006) BioSystems
- Faeder, Blinov, Goldstein & Hlavacek (2005) Syst. Biol.
- Faeder, Hlavacek, Reischl, Blinov, Metzger, Redondo, Wofsy & Goldstein (2003) J. Immunol.
- Goldstein, Faeder, Hlavacek, Blinov, Redondo & Wofsy (2002) Mol. Immunol.
Using Data in Modeling
Biological research is becoming increasingly complex and data-rich, with multiple public databases providing a variety of resources: hundreds of thousands of substances and interactions, hundreds of ready to use models, controlled terms for locations and reaction types, links to reference materials (data and/or publications), etc. Mathematical modeling should take advantage of this complex data and create quantitative, testable predictions based on the current state of knowledge.
- Blinov, Schaff, Ruebenacker, Wei, Vasilescu, Gao, Morgan, Ye, Lakshminarayana, Moraru, & Loew (2013) Bioinformatics
- Ruebenacker & Blinov (2010) Theory Biosci.
- Demir, ..., Blinov, ... (2010) Nature Biotechnology
- Blinov, Ruebenacker, Schaff, Moraru (2010) Lecture Notes in Bionformatics
- Ruebenacker, Moraru, Schaff & Blinov (2009 ) IET Systems Biology
- Igarashi, Heureux, Doctor, Talwar, Gramatikova, Gramatikoff, Zhang, Blinov, Ibragimova, Boyd, Ratnikov, Cieplak, Godzik, Smith, Osterman, & Eroshkin (2009). Nucleic Acids Res.
- Ruebenacker, Moraru, Schaff & Blinov (2007) Proc. 2007 IEEE BIBM
Models Storage and Visualization
A large mechanistic model (accounting for many species and activities and interactions among domains of biomolecules) is very difficult to store, visualize, or modify. The standard way of storage is in electronic exchange formats (e.g., SBML). SBML file specifies each of individual species and interactions, but carries no information about domains of proteins and composition of multi-protein species. Simulation and visualization tools (such as CellDesigner) display each species and interactions, making representation very cluttered. An alternative is specifying (storage, visualization) of key features of the system, sufficient to restore the complete model. One way is Molecular Interaction Maps (MIM). Rules provide another very convenient way of protein-protein interaction data representation. Even more important, rules can serve as templates, allowing large models to be composed of simpler models describing known interactions. Integration of rules, MIMs and SBML is one of my projects.