Prediction of genetic biomarkers from RNA-SEQ datasets of prostate cancer

A type of cancer known as prostate cancer (PC) manifests itself in the tissues of a man’s prostate gland. A walnut-sized gland between the penis and the bladder is prostate cancer! Prostate cancer symptoms include difficulty urinating, a decrease in the force of the urine stream, blood in the urine and semen, bone discomfort, weight loss, and erectile dysfunction. Four factors are associated with Prostate cancer: advanced age, race, family history, and obesity. Men aged 50 and older frequently develop prostate cancer. This project used different machine learning models to analyze the data gathered. The machine learning that was considered were: Random Forest, K-Nearest Neighbors, Gradient Boosting, Linear Discriminant Analysis (LDA), Support Vector Machines, and Naïve Bayes Classification. Prostate cancer, as well as ovarian and breast cancer, have all been related to BRCA1 and BRCA2 mutations. When a man has this mutation and is 65 or older, the patient is in danger of progressing to a more severe stage of the illness. As PC and breast cancer are connected to BRCA1 and BRCA2, patients with a family history of PC and breast cancer have a 60% likelihood of acquiring PC. More than 200 genes were examined, and HOXB13 G84E was shown to be a high-risk gene in PC patients. Doctors look for biomarkers to assist in detecting various malignancies and disorders. Biomarkers can be used to examine how a patient’s body reacts to a course of treatment for their ailment. It was challenging to decide which model would produce the best performance while choosing. Five A variety of machine learning models were chosen. The association between CACNG3 and ZNF263 in the prostate cancer setting raises the possibility of a functional connection between these two genes. It might suggest a joint function in encouraging carcinogenesis or regulating critical cancer-related pathways. The RandomForest ML model was the most effective one for this assignment. A machine learning system called random forest consists of various decision trees. Its primary goal is to solve all classification and regression problems. The genes KRT33A and CACNG3 were discovered to have a significant association with prostate cancer and disease development in this study. Overall, this thorough analysis highlights the complex dynamics of prostate cancer, covering its causes, detection techniques, genetic risk factors, and creative machine-learning applications. Keywords: Prostate Cancer, Genetic Biomaker, RNA-Seq, Machine Learning.