RNAfcg: RNA Flexibility\nPrediction Based on Topological\nCentrality and Global Features

The dynamics of RNAs are related intimately to their\nfunctions.\nMolecular flexibility, as a starting point for understanding their\ndynamics, has been utilized to predict many characteristics associated\nwith their functions. Since the experimental measurement methods are\ntime-consuming and labor-intensive, it is urgently needed to develop\nreliable theoretical methods to predict RNA flexibility. In this work,\nwe develop an effective machine learning method, RNAfcg, to predict\nRNA flexibility, where the Random Forest (RF) is trained by features\nincluding the topological centralities, flexibility–rigidity\nindex, and global characteristics first introduced by us, as well\nas some traditional sequence and structural features. The analyses\nshow that the three types of features introduced first have significant\ncontributions to RNA flexibility prediction, among which the topological\ntype contributes the most, which indicates the importance of structural\ntopology in determining RNA flexibility. The performance comparison\nindicates that RNAfcg outperforms the state-of-the-art machine learning\nmethods and the commonly used Gaussian Network Model (GNM) models,\nachieving a much higher Pearson correlation coefficient (PCC) of 0.6619\non the test data set. This work is helpful for understanding RNA dynamics\nand can be used to predict RNA function information. The source code\nis available at https://github.com/ChunhuaLab/RNAfcg/.