Aminormotiffinder - a graph grammar based tool to effectively search a minor motifs in 3D RNA molecules

RNA Motifs are three dimensional folds that play important role in RNA folding and its interaction with other molecules. They basically have modular structure and are composed of conserved building blocks dependent upon the sequence. Their automated in silico identification remains a challenging task. Existing motif identification tools does not correctly identify motifs with large structure variations. Here a “graph rewriting” based method is proposed to identify motifs in real three dimensional structures. The unique encoding of A Minor Searcher takes into consideration the non canonical base pairs and also multipairing of RNA structural motifs. The accuracy is demonstrated by correctly predicting A minor motifs across many PDB files with zero false positives. There is a huge demand of a good well developed RNA Motif identification algorithm that would successfully identify both canonical / non canonical and isomorphic motifs. In this thesis, a novel encoding algorithm is demonstrated that successfully identifies RNA A Minor Motifs from 3D RNAs. The algorithm encodes the three dimensional RNA Data into one dimension without losing any tertiary information during the transition. A Minor motif is then searched in this one dimensional string using exhaustive search technique with linear time complexity. The efficiency is demonstrated by the comparison of AMinorSearcher with benchmark tool FR3D. FR3D lacked in both precision and recall while AMinorSearcher did not