Title: Statistical Inference and Model Checking in Systems Biology
Abstract: Advances in
genomic technologies have revolutionized molecular biology, enabling
biologists to simultaneously probe thousands of genes or proteins in a
single experiment in a precise, fast and cost-effective way. Given the
high-dimensional data, the big challenge facing the biologists is how
to reconstruct the biological network. Statistical inference algorithms
play an important role in learning the biological network from
high-dimensional data. Different structure learning algorithms might
infer different optimal networks from same data due to the limited
number of experimental replications and high noise. Without
verification, the inferred optimal network cannot help us correctly
understand the cellular mechanisms. Model Checking is a formal
verification technique for the complex systems, including hardware
(e.g., CPU) and software (e.g., aerospace control software). It is the
process of determining whether or not a model satisfies a temporal
logic formula describing a desired behavior of the system. To overcome
the drawback of traditional techniques, we have integrated the
statistical learning methods with model checking techniques in a
unified framework to automatically infer and verify biological networks
from microarray data.