BayesMiRNAfind gene prediction Web Server

Inventors: Malik Yousef, Louise C. Showe and Michael K. Showe

Tech ID: SL06001

Description: The technology combines multi-species genomic data allowing for the identification of new miRNAs that does not rely on sequence conservation with known miRNAs. It is also useful for the confirmation of the status of a cloned small RNA sequence or to determine whether a genomic sequence can potentially code for a miRNA gene. A score is provided that assesses how well a sequence is classified. This technology is based on machine learning using a Naïve Bayes classifier.

Applications and Advantages: This program may be useful as a tool for microRNA gene predictions. The program does not depend on sequence homology. It generates a model automatically using rules based on miRNA gene structure thus allowing prediction of novel miRNAs that may not share sequence conservation with already known miRNAs as well as conserved species. A comparative analysis is used to reduce the false positive discovery rate. New microRNA species may be useful in cancer diagnosis or treatment. In tests on known miRNAs, BayesMiRNAfind outperforms related miRNA discovery tools

Licensing Opportunity: This technology is available for non-exclusive licensing to companies involved in the development and sale of microRNA discovery and analysis platforms.

Key Words: miRNA, microRNA, microRNA gene prediction, BayesMiRNAfind, “regulation of gene expression”, “regulation of messenger RNA”, “cancer diagnosis”

Contact:

Meryle J. Melnicoff
Director, Business Development
The Wistar Institute
3601 Spruce Street
Philadelphia, PA 19104
Phone: (215) 898-0049
Fax: (215) 573-2456
melnicoff@wistar.org

Last updated: Aug-02