AUTHOR=Han Ke , Wang Miao , Zhang Lei , Wang Ying , Guo Mian , Zhao Ming , Zhao Qian , Zhang Yu , Zeng Nianyin , Wang Chunyu TITLE=Predicting Ion Channels Genes and Their Types With Machine Learning Techniques JOURNAL=Frontiers in Genetics VOLUME=Volume 10 - 2019 YEAR=2019 URL=https://www.frontiersin.org/journals/genetics/articles/10.3389/fgene.2019.00399 DOI=10.3389/fgene.2019.00399 ISSN=1664-8021 ABSTRACT=Motivation: As a member of the gene family, the number of ion channels is increasing rapidly. Many of them are associated to diseases. Therefore, they are targets for more than 700 drugs. Discovery of new ion channels is facilitated by computational methods that predict the ion channels and their types from the protein sequences. Methods: We extracted the physicochemical properties of coding gene translation sequences by SVMProt and the k-skip-n-gram methods to get the feature vectors of ion channels, and obtain the 188-dimensional and the 400-dimensional features, respectively. The 188-dimensional features and the 400-dimensional features were combined to obtain the 588-dimensional features. We then employed the Maximum-Relevance-Maximum-Distance method to reduce the dimensions of the 588-dimensional features. Finally, the support vector machine and the random forest methods are used to build the prediction models to evaluate the classification effect. Results: Different methods are employed to extract the various feature vectors, and after effective dimensionality reduction, different classifiers are used to classify the ion channels. We extract the ion channel data from the Universal Protein Resource (UniProt, http://www.uniprot.org/) and the Ligand-Gated Ion Channel databases (http://www.ebi.ac.uk/compneur-srv/LGICdb/LGICdb.php), and then verify the performance of the classifiers after screening. In summary, this study can provide new information for the research and development of drugs.