Genomics at the Nexus of AI, Computer Vision, and Machine Learning

12Comparison of Machine Learning and Deep Learning Algorithms for Diabetes Prediction Using DNA Sequences

Gagandeep Kaur^1*, Poorva Agrawal¹, Latika Pinjarkar¹, Rutuja Patil², Suhashini Chaurasia³ and Seema Patil⁴

¹Symbiosis Institute of Technology, Nagpur Campus, Symbiosis International (Deemed University), Pune, India

²Vishwakarma Institute of Information Technology, Pune, India

³S.S. Maniar College of Computer and Management, Nagpur, India

⁴Symbiosis Institute of Technology, Pune, Symbiosis International (Deemed University), Pune, India

Abstract

The primary goal of the proposed work is to predict diabetes from deoxyribonucleic acid (DNA) sequence data by constructing a robust classification model. The evaluation of the model’s efficacy was conducted using the GenBank DNA dataset. The data underwent preprocessing, and the resulting feature vector, generated through the one-hot vector encoding technique, was supplied to diverse machine learning (ML)- and deep learning (DL)-based methods. We used advanced deep learning techniques, such as convolutional neural networks (CNN) and long termshort memory (LSTM), along with the random forest classifier, to predict diabetes. We fixed the size of K-mer at 4. According to our findings, LSTM outperformed random forest and CNN in classifying DNA sequences to predict diabetes. In this study, a methodological and flexible approach is presented for categorizing DNA sequence data by diabetes type and non-diabetic type.

Keywords: Diabetes ...