PhD defence: Sabarinathan Radhakrishnan
2014-01-31: Impact of mutations in structured RNAs. The defence will take place on January 31st at 13:15 in auditorium A2-70.03, Thorvaldsensvej 40, ground floor, University of Copenhagen, LIFE, Frederiksberg.
Everybody is welcome. Registration is not necessary.
After the defence a reception will be held at 16:30 on Grønnegårdsvej 7 (bldg. 1-04), 1st floor, Library/M106.
- Associate Professor Jakob Hull Havgaard, Department of Veterinary Clinical and Animal science, University of Copenhagen (Chairman)
- Associate Professor Christian Nørgaard Storm Pedersen, Bioinformatics Research Center, Faculty of Science, Aarhus University
- Associate Professor Danny Barash, Department of Computer Science, Ben-Gurion University
Professor Jan Gorodkin, Center for non-coding RNA in Technology and Health, IKVH, KU.
The essential structural features of many functional non-coding RNAs and cis-acting regulatory elements in messenger RNA are in turn determined by their primary sequence. The genetic variations, such as single nucleotide polymorphisms and mutations, in those RNA sequences can alter their structure and thereby affect the function of RNA. In recent years, several computational methods have been developed to predict the mutation effects on RNA secondary structure; however, they have the following limitations: a) inability to explicitly detect mutation effect on local RNA secondary structure, and b) being computationally expensive for genome-wide applications.
In this study, we have developed a method, called RNAsnp, to aid the prediction of mutation induced structural changes in local regions of RNA secondary structure. The secondary structure can be predicted globally or locally in a given sequence, using the RNA folding algorithms implemented in Vienna RNA package. The confidence in the mutation effects are measured in terms of empirical P-values, which are derived from extensive pre-computed tables of the distribution of mutation effects predicted on random sequences of various lengths and GC contents. The combination of local structure prediction and the pre-computed tables makes RNAsnp sufficiently fast for genome-wide applications. The performance of RNAsnp was evaluated using already known structure-disruptive mutations reported in the literature. Furthermore, applying RNAsnp to a set of 514 human disease-associated mutations resulted in 54 predictions with significant structural effect on RNA secondary structure (P -value < 0.1). We also developed a web sever that provides an interface to select the different features of RNAsnp and visualize the output with graphical representation. The web server is freely available at http://rth.dk/resources/rnasnp.
Further, we show the application of RNAsnp on a large variation data, from transcriptome-wide sequencing of non-small cell lung cancer, to screen putative structure-disruptive variants. Finally, we present a method (RNAbound) based on the strategy for detection of locally structured regions, previously implemented in RNAsnp, to detect boundaries of RNA structure in the multiple genome alignment. We found that the application of RNAbound together with RNAz enhance the prediction of number of high confidence structured RNAs (P > 0.9) by 1.5 fold compared to the RNAz screen on fixed size sliding windows.