Related Experiment Videos
Bayesian assignment of gene ontology terms to gene expression experiments
1Department of Biotechnology, BOKU University, Muthgasse 18, 1190 Vienna. peter.sykacek@boku.ac.at
Bioinformatics (Oxford, England)
|September 11, 2012
Summary
This study introduces a probabilistic model for inferring Gene Ontology (GO) terms from gene expression data. This approach enhances biological interpretation, especially for sparsely annotated genes and uncertain gene activity, improving upon statistical methods.
Area of Science:
- Bioinformatics
- Computational Biology
- Genomics
Background:
- Gene expression assays enable genome-scale molecular mechanism analysis.
- Interpreting gene lists from expression data is challenging.
- Gene Ontology (GO) term inference is a key method for high-level biological interpretation.
Purpose of the Study:
- To propose a probabilistic model for GO term inference from gene expression data.
- To enhance the biological interpretation of gene expression analyses.
- To provide a probabilistic GO term assignment.
Main Methods:
- Developed a probabilistic model integrating gene annotations and posterior probabilities of gene activity.
- Combined indicator probabilities from Bayesian expression data analysis.
- Applied the model to synthetic and microarray data.
Main Results:
- The probabilistic model effectively infers GO terms.
- Outperforms statistical test-based approaches, particularly for sparsely annotated GO terms.
- Demonstrates advantages in situations with high uncertainty in gene activity.
Conclusions:
- The proposed probabilistic model offers improved GO term inference.
- Easily adaptable for inferring other high-level biological assignments like pathways.
- Provides a robust method for interpreting complex gene expression data.