Predictive Diabetes Mellitus From DNA Sequences Using Deep Learning

Diabetes is a chronic metabolic disorder characterized by elevated blood sugar levels. It manifests in different forms, with type 1 and type 2 being the most prevalent. Type 1 diabetes results from the autoimmune destruction of insulin-producing cells, whereas type 2 diabetes primarily stems from insulin resistance. Despite advancements in treatment, accurate detection and prediction of diabetes remain challenging. Early diagnosis is crucial for effective management and prevention of complications. Another obstacle lies in interpreting vast amounts of health data, including DNA sequencing, which poses difficulties for healthcare professionals in identifying relevant patterns and associations. Artificial intelligence (AI) holds promise in healthcare by developing and training deep learning algorithms to analyze health data and DNA sequences. The research paper focuses on applying both Convolutional Neural Networks (CNNs) algorithm, in addition to Long Short-Term Memory (LSTM) algorithm for predicting types of diabetes based on DNA sequencing. The study aims to leverage the power of CNN and LSTM, known for their proficiency in analyzing image and sequence data, to accurately classify diabetes types. The experimental results of the proposed CNN-LSTM model showcased remarkable performance, achieving a recorded accuracy of 100% on a labeled dataset that included DNA sequencing and corresponding diabetes types. The model's evaluation encompassed several metrics, including accuracy, recall, precision, and the F1 score.