H-DNA Areas Recognition Using Machine Learning Methods

A lot of discoveries have been made in genetic engineering in the last decades. The rapid growth in such areas as machine learning and bioinformatics prompted the corresponded growth in creating vast range of models which are able to find patterns in DNA structures and even to generate their different chains. However, considering the fact that CNN (convolutional neural network) models are the best in that area, the problem of selecting prevailed factors in model decision-making has become complicated. The main goal of this paper is to solve the H-DNA segments recognition problem in human genome using different ML methods. In particular, we will obtain the CNN model with the high recognition accuracy in order to define the significant features of chain parts. This model will respond why a chain part is exactly an H-DNA.