Reconstruction of gene networks using Bayesian learning and manipulation experiments

Authors:
Iosifina Pournara;Lorenz Wernisch
Affiliations:
Department of Crystallography, Birkbeck College, University of London, Malet Street, London, WC1E 7HX, UK;Department of Crystallography, Birkbeck College, University of London, Malet Street, London, WC1E 7HX, UK
Venue:
Bioinformatics
Year:
2004

Citing 0
Cited 6

Inferring biomolecular interaction networks based on convex optimization

Computational Biology and Chemistry
Constructing gene networks using variational bayesian variable selection

Artificial Life
Learning the structure of dynamic Bayesian networks from time series and steady state measurements

Machine Learning
Local Causal and Markov Blanket Induction for Causal Discovery and Feature Selection for Classification Part I: Algorithms and Empirical Evaluation

The Journal of Machine Learning Research
A Swarm Intelligence Framework for Reconstructing Gene Networks: Searching for Biologically Plausible Architectures

IEEE/ACM Transactions on Computational Biology and Bioinformatics (TCBB)
Inferring gene interaction networks from ISH images via kernelized graphical models

ECCV'12 Proceedings of the 12th European conference on Computer Vision - Volume Part VI

Quantified Score

Hi-index	3.84

Visualization

Abstract

Motivation: The analysis of high-throughput experimental data, for example from microarray experiments, is currently seen as a promising way of finding regulatory relationships between genes. Bayesian networks have been suggested for learning gene regulatory networks from observational data. Not all causal relationships can be inferred from correlation data alone. Often several equivalent but different directed graphs explain the data equally well. Intervention experiments where genes are manipulated can help to narrow down the range of possible networks. Results: We describe an active learning algorithm that suggests an optimized sequence of intervention experiments. Simulation experiments show that our selection scheme is better than an unguided choice of interventions in learning the correct network and compares favorably in running time and results with methods based on value of information calculations. Availability: Algorithms are available from the authors on request.