Comparison of evolutionary algorithms in gene regulatory network model inference
Sîrbu, Alina and Ruskin, Heather J. and Crane, Martin (2010) Comparison of evolutionary algorithms in gene regulatory network model inference. BMC Bioinformatics, 11 (59). pp. 1471-2205. ISSN 1471-2105
Full text available as:
Background: The evolution of high throughput technologies that measure gene expression levels has created a
data base for inferring GRNs (a process also known as reverse engineering of GRNs). However, the nature of
these data has made this process very di±cult. At the moment, several methods of discovering qualitative
causal relationships between genes with high accuracy from microarray data exist, but large scale quantitative
analysis on real biological datasets cannot be performed, to date, as existing approaches are not suitable for real
microarray data which are noisy and insu±cient.
Results: This paper performs an analysis of several existing evolutionary algorithms for quantitative gene
regulatory network modelling. The aim is to present the techniques used and o®er a comprehensive comparison
of approaches, under a common framework. Algorithms are applied to both synthetic and real gene expression
data from DNA microarrays, and ability to reproduce biological behaviour, scalability and robustness to noise are assessed and compared.
Conclusions: Presented is a comparison framework for assessment of evolutionary algorithms, used to infer gene
regulatory networks. Promising methods are identi¯ed and a platform for development of appropriate model
formalisms is established.
Archive Staff Only: edit this record