DNA Sequence Classification Using RRCNN and LSTM

Advancements in genomics have led to an exponential increase in the availability of DNA sequence data, offering a rich source of information for various biomedical applications, including disease prediction, functional annotation, and evolutionary analysis. Efficient and accurate classification of DNA sequences is paramount to unlocking the hidden knowledge within these vast datasets. This wealth of genetic information presents a remarkable opportunity for advancing our understanding of viral pathogens such as SARS, and MERS. Effective classification of these viral DNA sequences is crucial for epidemiological studies, drug development, and vaccine design. This paper introduces a novel approach, employing Recurrent Residual Convolutional Neural Networks (CNN) with Long Short-Term Memory (LSTM) networks, for the classification of virus DNA sequences. Our proposed model, RRCNN-LSTM, integrates the strengths of CNNs in extracting local sequence features and LSTMs in capturing long-range dependencies within the sequences. The utilization of residual connections in the CNN component enhances the gradient flow during training, facilitating the modeling of intricate sequence patterns. Additionally, the LSTM component effectively handles variable-length sequences, making the RRCNN-LSTM model versatile for a wide range of DNA sequence classification tasks. This proposed approach gave a remarkable accuracy of 96.27%, which helps in classifying the viruses.