Machine Learning Meets Big Data: An Overview of Diagnostic and Prognostic Prediction for Cancer

The past decades have witnessed the rapid development on biological technology and artificial intelligence (AI). Large amounts of omics data have been produced and accumulated by high-throughput sequencing technology. The accumulation of multi-omics data and the innovation of machine-learning provide a variety of resources for diagnostic and prognostic prediction of cancer, which efficiently enhance clinical decision making, improve prognosis, and accelerate the progress of precision medicine. In this review, first we reviewed the common omics data. Second, we briefly introduced several popular machine learning methods. Thirdly, we systematically summarized the substantial achievements obtained in a number of prediction studies of cancer, including breast cancer (BRCA), kidney renal clear cell carcinoma (KIRC), glioblastoma multiforme (GBM). Lastly, it is promising that the advanced AI technologies such as deep learning will be widely applied to facilitate diagnosis and treatment for cancer patients as massive omics data are efficiently and accurately measured.