Deep Learning Techniques in Cancer Prediction Using Genomic Profiles

According to World Health Organization (WHO), cancer is the second major reason of death followed by cardiovascular diseases. Early diagnosis of cancer is essential to offer correct treatment to the patient. Traditional histology based examination is used as the gold standard for cancer identification but the accuracy of the result generated is questionable and it may vary amongst pathologists. Next-Generation Sequencing (NGS) technologies have produced a huge amount of cancer genomic data publicly available which elevates an idea of the identification of candidate genes contributing to uncontrolled cell growth resulting in cancer. Analyzing gene expression data is crucial to find out these harmful mutations and to avoid further consequences. Therefore there is a need for learning methods to explore this data for discovery of target genes, accurate disease identification and drug discovery. Despite research in cancer detection using genome sequencing, there exists a need to improve accuracy and find out driver mutations in genes. This paper presents a systematic literature survey of in-depth learning techniques used to predict cancer using genome sequencing. The system can be used by medical practitioners to have timely diagnosis & prognosis of diseases like cancer.