Floating search methods in feature selection
Pattern Recognition Letters
Pattern Classification (2nd Edition)
Pattern Classification (2nd Edition)
Information-theoretic inference of large transcriptional regulatory networks
EURASIP Journal on Bioinformatics and Systems Biology
Inferring Connectivity of Genetic Regulatory Networks Using Information-Theoretic Criteria
IEEE/ACM Transactions on Computational Biology and Bioinformatics (TCBB)
AGN Simulation and Validation Model
BSB '08 Proceedings of the 3rd Brazilian symposium on Bioinformatics: Advances in Bioinformatics and Computational Biology
| Hi-index | 0.00 |
An important problem in the bioinformatics field is to understand how genes are regulated and interact through gene networks. This knowledge can be helpful for many applications, such as disease treatment design and drugs creation purposes. For this reason, it is very important to uncover the functional relationship among genes and then to construct the gene regulatory network (GRN) from temporal expression data. However, this task usually involves data with a large number of variables and small number of observations. In this way, there is a strong motivation to use pattern recognition and dimensionality reduction approaches. In particular, feature selection is specially important in order to select the most important predictor genes that can explain some phenomena associated with the target genes. This work presents a first study about the sensibility of entropy methods regarding the entropy functional form, applied to the problem of topology recovery of GRNs. The generalized entropy proposed by Tsallis is used to study this sensibility. The inference process is based on a feature selection approach, which is applied to simulated temporal expression data generated by an artificial gene network (AGN) model. The inferred GRNs are validated in terms of global network measures. Some interesting conclusions can be drawn from the experimental results, as reported for the first time in the present paper.