Prediction of Ion Channels Genes and Types

Creative Bioarray is committed to developing a range of machine learning (ML) algorithms to help clients accurately predict and classify ion channels. Our bioinformatics analysis services will help you understand the pathological mechanism of ion channel gene mutation causing related diseases.

Introduction

Ion channels exist in many tissues of organisms and directly or indirectly lead to many diseases, such as retinal diseases, arrhythmias migraine, renal cysts, deafness, and epilepsy. Studies have shown that about 19% of human protein drug targets belong to the ion channel family. With the rapid growth of proteomic data and the progress in the study of ion channel structure and function, more and more researchers pay attention to the early prediction and identification of ion channel genes and types.

The interest in drugs targeting ion channels and the need to expand ion channel protein annotation have driven the development of high-throughput computational tools that allow the prediction of ion channels and their types through protein sequences. The development of these bioinformatics technologies has accelerated the research progress in various fields such as drug repurposing, disease understanding, and drug discovery planning. Researchers are now increasingly using high-throughput computational tools to predict and classify ion channels.

Fig. 1 Two alternative ways to predict the ion channels and ion channel types. (Gao, 2019)

Our Services

ML is a powerful tool for extracting feature vectors of a variety of ion channels. We have developed multiple computational methods based on ML algorithms to predict the genes of ion channels, as well as their types using protein sequences. Currently, we mainly apply different methods including SVMProt and k-skip-n-gram methods to extract various feature vectors, and build predictive models to classify ion channels after effective dimensionality reduction. Our ion channel gene and type prediction services include the following steps:

• Feature extraction of samples by two methods, including SVMProt 188-D feature extraction and k-skip-n-gram 400-D feature extraction.
• Feature selection. Generation of sub-features with low redundancy and strong correlation through maximum-relevance-maximum-distance (MRMD)-based dimensionality reduction method to achieve more accurate classification results.
• Comparative assessment of the ion channel predictors.
• Estimation of the prediction model accuracy.

Furthermore, we constantly improve and develop new prediction and classification models to adapt to the rapid increase of ion channel types and data.

Applications

• Study on the roles of ion channels in diseases
• Drug development

Creative Bioarray has accumulated rich experience in extracting ion channel features with new features to help clients predict the gene and type of ion channels. Our research team has unparalleled expertise in computational intelligence, ML, and ion channels to ensure efficient scientific service to you. If you are interested in our services, please feel free to contact us for more details.

Reference

1. Gao, J.; et al. Prediction of ion channels and their types from protein sequences: Comprehensive review and comparative assessment. Current Drug Targets, 2019, 20(5): 579-592.