Machine Learning based Identification of Cancer Related Mirna and Gene Biomarkers

Abstract

 Cancer is still a major killer, thus finding effective biomarkers and making correct diagnoses requires cutting-edge computational methods. Using data from The Cancer Genome Atlas (TCGA), this study investigates the predictive power of deep learning and machine learning models in finding genes and miRNAs unique to cancer. Supervised learning models including Random Forest, Support Vector Machine (SVM), and Gradient Boosting were used in a thorough examination, along with deep learning architectures like CNNs and LSTM networks. To further understand the intricate relationships between miRNAs and genes, we used graph-based approaches. Using important performance indicators including accuracy, precision, recall, F1 score, and Area under the Curve (AUC), the models were assessed using K-fold cross-validation. For the purpose of differentiating cancer-specific biomarkers, the findings showed that deep learning models, especially CNN (AUC = 0.95) and LSTM (AUC = 0.93), performed better than conventional machine learning methods. By providing data-driven approaches for early identification and individualized treatment regimens, this work highlights the potential of deep learning to advance precision oncology.