AUTHOR=Zhang Xu , Liu Yiwei , Wang Yaming , Zhang Liang , Feng Lin , Jin Bo , Zhang Hongzhe 

TITLE=Multistage Combination Classifier Augmented Model for Protein Secondary Structure Prediction

JOURNAL=Frontiers in Genetics

VOLUME=Volume 13 - 2022

YEAR=2022

URL=https://www.frontiersin.org/journals/genetics/articles/10.3389/fgene.2022.769828

DOI=10.3389/fgene.2022.769828

ISSN=1664-8021

ABSTRACT=With the accomplishment of gene sequencing works, obtaining protein structure information by pattern recognition and machine learning is an important task of Bioinformatics. In the field of Bioinformatics, protein secondary structure prediction plays an important role in many biological aspects. Accurate secondary structure prediction not only can help to determine the three-dimensional structure of protein, but also provide important information for protein function annotation. Considering the physical experiment based protein secondary structure prediction methods are time-consuming and expensive, some deep learning methods are proposed to try to solve the classification task. However, few researchers try to solve the protein secondary structure prediction task from the perspective of model learning or training methods. In this paper, a method called  Multistage Combination Classifier Augmented Model (MCCM) is proposed to solve the protein secondary structure prediction task in the perspective of model learning method. At the core is that two classifier branches attached to a backbone network are used to learn easier and harder samples, which can increase the loss value of the harder samples and induce model to pay more attention to them and improve the classification performance. The experimental results on the publicly available benchmark CB513 data set outperforms other state-of-the-art models.