Inferring Causation in Yeast Gene Association Networks With Kernel Logistic Regression

Area of Science:

• Bioinformatics
• Systems Biology
• Computational Biology

Background:

• Inferring gene-gene associations is crucial in bioinformatics.
• Existing tools often lack directionality and reaction type information.
• Causal gene networks require methods that consider these relationships.

Purpose of the Study:

• To construct a causal gene co-expression network.
• To identify transcription factors within gene pairs.
• To improve prediction accuracy using machine learning.

Main Methods:

• Utilized microarray expression data from Saccharomyces cerevisiae.
• Employed a machine learning technique based on logistic regression.
• Classified gene pairs into connected or nonconnected based on correlation.

Main Results:

• Successfully constructed a causal gene co-expression network.
• Achieved high performance in predicting gene relationships.
• Demonstrated effectiveness in identifying transcription factors.

Conclusions:

• The logistic regression model effectively addresses network sparsity.
• The proposed system enhances the accuracy of predicting gene associations.
• This method provides a robust framework for yeast regulatory network analysis.