EPEL

Instruction

Effect prediction of cancer-specific synonymous mutations

Features

DNA shape features, physicochemical properties,and deep learning-derived features based chemical molecule

Application

Facilitating the exploration of the mechanisms and the screening of potential driver sSNVs in cancer

performance

EPEL demonstrates superior performance compared to the exist state-of-the-art methods

What is EPEL ? What is EPEL ? EPEL was developed based an ensemble learning method that utilizes sequence representation to predict driver sSNVs. We first investigate the contribution of sequence-based features, including DNA shape, physicochemical properties based on encoded single nucleotide, and deep learning-derived features from pre-trained chemical molecule language models. Secondly, we propose EPEL, an effect predictor of synonymous mutations, employing ensemble learning. EPEL combines five tree-based models and optimizes feature selection to enhance predictive accuracy. Notably, the application of DNA shape features and deep learning-derived features based chemical molecule represents a pioneer effect in assessing the impact of synonymous mutations. Compared to the exist state-of-the-art methods, EPEL demonstrates superior performance on independent test set. The EPEL code and corresponding data are available at https://github.com/maxcine-cloud/EPEL.