Related Experiment Videos
Genes, themes and microarrays: using information retrieval for large-scale gene analysis
H Shatkay1, S Edwards, W J Wilbur
1National Center for Biotechnology Information, NLM, NIH, Bethesda, Maryland 20984, USA. shatkay@ncbi.nlm.nih.gov
Summary
This study introduces a novel literature-based method to uncover genome-wide gene functions, overcoming limitations of traditional clustering in DNA microarray analysis. The approach effectively translates text themes into functional gene connections.
Area of Science:
- Bioinformatics
- Computational Biology
- Genomics
Background:
- DNA microarray experiments generate vast datasets, posing significant analysis challenges.
- Current genome-wide expression analysis often uses clustering, which identifies gene patterns but not underlying biological mechanisms.
Purpose of the Study:
- To develop a new method for establishing genome-wide functional relationships among genes using scientific literature.
- To address the limitations of clustering methods in explaining biological mechanisms.
Main Methods:
- A novel approach utilizing literature-based analysis to identify functional gene relationships.
- Employing a similarity-based search within document space to reveal coherent themes.
- Translating content-based relationships from abstracts into functional gene connections.
Main Results:
- Preliminary experiments applied the algorithm to a database of yeast gene documents.
- The method successfully identified functional connections among genes based on literature themes.
- Results demonstrated effectiveness when compared against established yeast gene functions.
Conclusions:
- The developed literature-mining approach offers a powerful tool for genome-wide functional gene analysis.
- This method complements traditional expression data analysis by elucidating biological mechanisms.
- The approach shows promise for advancing our understanding of gene functions across various organisms.