3D RNA Graph Representation Methods for Classification of RNA Molecules Using Graph Kernel and Graph Neural Network Methods

Authors : Enes Algül

Pages : 919-934

Doi:10.53433/yyufbed.1256154

View : 60 | Download : 66

Publication Date : 2023-12-29

Article Type : Research Paper

Abstract :Ribonucleic acids (RNAs) are nucleic acid types with 1D/2D/3D structural shapes and are essential for sustaining life. These structural shapes of the RNAs are highly correlated with their functions. While the primary and secondary structures of RNA have been extensively studied, the tertiary structure has received relatively less attention. In this article, we present novel approaches for representing 3D RNA structures as graph data, employing geometric measurements such as Base position, Square root velocity function (SRVF), Arc length, and Curvature. Then, we utilise kernel methods and neural network methods to predict RNA functions. Our findings demonstrate the effectiveness of these methodologies in unraveling the functional attributes of RNA molecules, thus enriching our understanding of their complex biological significance.
Keywords : 3B RNA Graf Temsilleri, Geometrik Ölçümler, Graf Çekirdekleri, Graf Sınıflandırmaları, Graf Sinir Ağları