Applications of artificial intelligence in precision medicine

Precision medicine (PM) can be understood as an approach to the prevention and treatment of diseases through such development of diagnostics and therapy that takes into account information about particular genes and integrates... [ view full abstract ]

Precision medicine (PM) can be understood as an approach to the prevention and treatment of diseases through such development of diagnostics and therapy that takes into account information about particular genes and integrates clinical and molecular information, as well as the patient's environment and lifestyle. Advances in biological and medical technologies provide us with enormous amounts of data in the form of images, texts, numbers or multimedia messages. Genome sequencing, advanced biotechnology or various health sensors used by patients, with handheld devices, including smartphones, watches, etc., produce a huge amount of data. Learning from these data makes it easier to understand human health and diseases. Big data is so huge and complex that traditional data processing software is not enough to deal with it. Data of this type must be appropriately captured, stored, analyzed, searched, shared, transmitted, visualized and updated. Serious problems concern the privacy of medical information and the source of such data. In our times, the challenge is the problem of processing such data, which are usually noisy and often are in the form of streams coming in real-time. Machine learning techniques (ML) help in solving diagnostic and prognostic problems in various fields of medicine. ML is used to analyze the significance of clinical parameters and their combinations for the prognosis, e.g. prediction of disease progression, extraction of medical knowledge. All this in combination with individual clinical and molecular information, the environment and lifestyle factors, are a contribution to precision medicine. 

 The aim of this paper is to provide a review of recent machine learning techniques and some of the state-of-the-art applications used in precision medicine. We pay attention to why and where big data are and which methods have succeeded. We show examples of applications of machine learning, including classification of medical images, analysis of genomic sequences, as well as classification and prediction of protein structure. At the end, we present our point of view regarding future applications of computational intelligence in precision medicine.