A. Kuijper (Austria) and I. Havukkala (New Zealand)
Shape analysis, microRNA, mutual symmetry, matching, evaluation
In this work we investigate the use of mutual symmetric points in matching microRNA shapes. These 2D shapes, obtained by an RNA folding algorithm, are given as co ordinate points of 80-140 nucleotides long microRNA se quences containing A-U and C-G nucleotide pairs. The method of mutual symmetric points, exploiting global as pects, is well-suited for analysing these shapes that contain global symmetries. We discuss how a parameterisation can be derived from the coordinate point representation of these shapes. As this results in a data set with too many points, we in vestigate how reasonable subsets of this parameterisation can be used to compare microRNA shapes with respect to optimising both the computational effort and the search ac curacy and sensitivity. We conclude that taking a uniform step size parame terisation with about 3 times as many points as nucleotides performs best in view of computational efficiency and ef fectiveness. It should be taken into account whether shapes are aligned or not.
Important Links:
Go Back
