My long-term goals are to work as an independent investigator capable of solving the most challenging problems that arise in the statistical analysis of genetic data. I'm particularly interested in methods for study of gene interactions and regulatory pathways. My post-doctoral training will utilize the research environment at the division of Computational Biology of the Department of Biostatistics and Computational Biology at University of Rochester to enhance my understanding and knowledge of biological processes relevant to genetic analysis. The division of Computational Biology of this department is composed of 6 tenured (tenure-track) faculty members with a long record of methodological research in various areas on the boundary between statistics and biology. My current research is motivated by a collaboration with Dr. Land at the Biomedical Genetics Department. I would like to continue to collaborate with biologists and geneticists, expect that the proposed biological training will help me to became more efficient collaborator. The research goal of this proposal is to study the regulatory relations among genes based on gene expression data. We will address the theoretical and computational issues arising in the reconstruction of regulatory networks. In particular, we will propose theoretically based scores for network reconstruction in a frequentist and Bayesian framework. Searching for high scoring networks is computationally complex because of the super-exponential size of space of possible networks. We will analyze commonly used heuristic search algorithms, like Simulated annealing (SA) and Markov Chain Monte Carlo (MCMC). We will propose an implementation of the SA and MCMC algorithms that works on a reduced space of orders and thus increases the efficiency. Our approach will be validated by both simulated data as well as real biological data. We will explore biologically relevant questions based on regulatory networks such as, the optimal design of microarray experiments;the search for target genes for cancer therapy;the search for differentially expression of genes. The ultimate long term goal of this research is directed at understanding the basis of oncogenes cooperation and identification of malignant pathways. These advances will have direct effect on identification of targets for cancer cure.

Agency
National Institute of Health (NIH)
Institute
National Library of Medicine (NLM)
Type
Research Transition Award (R00)
Project #
5R00LM009477-04
Application #
8065511
Study Section
Biomedical Library and Informatics Review Committee (BLR)
Program Officer
Ye, Jane
Project Start
2009-01-01
Project End
2012-11-03
Budget Start
2011-01-01
Budget End
2012-11-03
Support Year
4
Fiscal Year
2011
Total Cost
$148,095
Indirect Cost
Name
University of Rochester
Department
Biostatistics & Other Math Sci
Type
Schools of Dentistry
DUNS #
041294109
City
Rochester
State
NY
Country
United States
Zip Code
14627
Balov, Nikolay (2011) A Gaussian Mixed Model for Learning Discrete Bayesian Networks. Stat Probab Lett 81:220-230