Artificial Intelligence Improves Genomic Medicine

Artificial Intelligence (AI) is the development and use of computing systems concerned with making machines work in an intelligent way. AI can fall under 2 different categories: (1) Narrow AI that focuses on performing a single specific task; or (2) General AI that refers to the concept of a machine that is able to perform different types of intelligent tasks and have ‘human’ reasoning.

Genomics is the study of all of the genetic material of an organism. Genomic medicine utilizes an individual’s genomic information to improve their clinical care and provide a more personalized approach for diagnostic or therapeutic decision making. 

A major goal of many large-scale genomic sequencing indicates, such as the 100,000 human genome project, is to advance genomic medicine by accelerating the identification and understanding of diseases and therapeutic associated genetic variants. Analyzing -omics and clinical data, that is essential for new biomedical discoveries for personalized medicine, has proven to have complex analytical and computational challenges for researchers. As a result, there is an increasing need to analyse large data sets with computational approaches and methods that can provide faster and more accurate analytical solutions.

Currently, existing programs of AI and machine learning (ML) have been used to treat cancer. Some programs have been used for annotating genomes- using ML methods to identify and classify specific sequences within each genome (for example: transcription start sites, promoters and splice sites). Understanding genomic features are important as they are related to important structural, functional and regulatory mechanisms, so accurate identification of genomic features are crucial to analyzing genomes clinically. Usually, the ML methods learn in order to detect specific patterns (ie-DNA sequences) in genomes. In recent years, there has been significant developments in methods to analyse genomes thoroughly. With these developments, more larger and detailed datasets have been generated that can aid in the discovery and prediction of genomic features. This is important to clinical genomics as it greatly impacts patient care and treatment. After incorporating deep learning algorithms, researchers are able to devise better strategies to improve the prediction of the effects of various genomic variants.

Figure 1:  the steps of genomic data processing, taken from Cambridge University

Figure 1: the steps of genomic data processing, taken from Cambridge University

The figure above illustrates the process of analyzing genomic data. First, a DNA sample must be obtained in order for the algorithm to determine its genomic sequence. After that, the genomic data will be processed and sorted according to their functions and genomic variants based on their likelihood to cause a disease. This is important to clinical genomics as it greatly impacts patient care and treatment. Once this is done, the data will be analysed and interpreted by the algorithm. The algorithm will compare the available data with pre-existing data from other databases to improve the accuracy of genomic function. This interpretation will be used in clinical reporting to improve patient care and treatment, or in research reporting to expand the genomic knowledge.

Thumbnail Credit: The Medical Futurist