PhD Application Talk: Computational prediction of spliced microRNAs in plants

Present computational detection methods focus on miRNA precursor structures rather than the mature sequences. However, the recent discovery of spliced miRNA sequences poses new challenges: Very large introns disrupting the short hairpin of the miRNA make these sequences unpredictable with current prediction methods.
I developed a plant-specific procedure to tackle this problem. The predictor consists of two phases. Potential stems are identified in phase 1. In phase 2, stems with sufficient complementarity to a given target gene are detected. By means of classifier systems putative intronic sequences are removed and the resulting miRNA candidates are checked for the capability of forming the distinctive hairpin structure. In my talk, I will briefly discuss and evaluate my prediction procedure on the rice and maize genomes based on spliced examples of the MIR444 family.
In summary, my predictor is the first bioinformatic approach for this new kind of miRNAs and, furthermore, still applicable to non-spliced miRNAs.