Stability of inferring gene regulatory structure with dynamic Bayesian networks

Authors:
Jagath C. Rajapakse;Iti Chaturvedi
Affiliations:
Bioinformatics Research Center, School of Computer Engineering, Nanyang Technological University, Singapore and Department of Biological Engineering, Massachusetts Institute of Technology, Cambrid ...;Bioinformatics Research Center, School of Computer Engineering, Nanyang Technological University, Singapore
Venue:
PRIB'11 Proceedings of the 6th IAPR international conference on Pattern recognition in bioinformatics
Year:
2011

Citing 2
Cited 0

Probabilistic Approximations of Signaling Pathway Dynamics

CMSB '09 Proceedings of the 7th International Conference on Computational Methods in Systems Biology
Learning the structure of dynamic probabilistic networks

UAI'98 Proceedings of the Fourteenth conference on Uncertainty in artificial intelligence

Quantified Score

Hi-index	0.00

Visualization

Abstract

Though a plethora of techniques have been used to build gene regulatory networks (GRN) from time-series gene expression data, stabilities of such techniques have not been studied. This paper investigates the stability of GRN built using dynamic Bayesian networks (DBN) by synthetically generating gene expression time-series. Assuming scale-free topologies, sample datasets are drawn from DBN to evaluate the stability of estimating the structure of GRN. Our experiments indicate although high accuracy can be achieved with equal number of time points to the number of genes in the network, the presence of large numbers of false positives and false negatives deteriorate the stability of building GRN. The stability could be improved by gathering gene expression at more time points. Interestingly, large networks required less number of time points (normalized to the size of the network) than small networks to achieve the same level stability.