דן ברש

This paper describes an efficient pattern recognition method for detecting vulnerable spots within an RNA sequence. Mutations in these spots may lead to a structural change that directly relates to a change in functionality. Previously, the concept was tried on RNA genetic control elements called 'riboswitches' and other known RNA switches. Here, the concept is extended to assist in planning in-vivo experiments in general, using a new tool that we have developed called RNAMute. We apply the package RNAMute on an RNA transcript that was shown experimentally to inactivate spectinomycin resistance in Escherichia coli by creating a library of point mutations using PCR and screening to locate those mutations. Our prediction, conducted independently of the known experimental results, succeeds in matching the inactivating point mutations that were obtained by the selection experiment. Validation of the method with data available from laboratory experiment supports its use as a general predictive tool.