Abstract
We have developed a technique for the reconstruction and validation of models of gene networks based on the gene expression profiles derived in the course of DNA microchip experiments or by the method of RNA molecules sequencing. A structural block diagram is presented of a stepwise process for determining optimal parameters of the algorithm for reconstruction of a gene network that meet the optimum network topology. We proposed a comprehensive estimation criterion of a gene network topology based on the Harrington desirability function that contains network topological parameters as constituent components. The maximum value of this criterion corresponds to the optimal topology of a gene network. A technique for the validation of models of gene networks is based on a ROC analysis whose implementation implies a comparative analysis of the character of relations between relevant genes in the network on the basis of the totality of genes and gene networks based on the obtained biclusters. Qualitative reconstruction of a gene network makes it possible to explore the nature of interaction between genes that determine the process of functioning of a biological organism at different stages of development of complex genetic diseases for the purpose of early diagnosis and correction of a given process. It was established that the gene network reconstructed based on the correlation output algorithm is more efficient in comparison with the gene network based on the algorithm ARACNE. The weighted average of relative validation criterion for the derived models based on the correlation output algorithm is significantly greater than the corresponding value when applying the algorithm of ARACNE. This fact indicates a higher degree of compliance with the character of relations between respective genes in the network based on the totality of genes and in the networks based on gene expression profiles in the obtained biclusters. Qualitative reconstruction of a gene network makes it possible to explore the character of development of a biological organism at the gene level, which creates preconditions for early diagnosis and adjustment of the development of different types of genetic diseases.
Highlights
Modern information processing systems in most cases are based on the use of analogies of functioning of biological mechanisms and processes occurring in living organisms
The maximum value of Harrington desirability index corresponds to the threshold coefficient, which makes it possible to obtain an optimal structure of a gene network based on the integrated analysis of topological parameters
We have developed a technique for the reconstruction of gene regulatory networks based on comprehensive application of network topological parameters and the generalized Harrington desirability index
Summary
Modern information processing systems in most cases are based on the use of analogies of functioning of biological mechanisms and processes occurring in living organisms. One of modern methods employed for the reconstruction of GRN is the identification of a network by comparing it with a known network of the relevant biological organism In this case, the criteria for comparison are the network topology and the existence of appropriate genes in a given network compared to a reference network. Reconstructed GRN makes it possible to explore the character of development of a biological organism at the gene level, which creates preconditions for early diagnosis and adjustment of the development of different types of genetic diseases. This fact demonstrates the relevance of research topic presented here
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have