The Intersection of AI, Genomics and Gene Editing
Artificial intelligence is known to have revolutionized a lot of industries.
From businesses accelerating the digital transformation to marketing managers developing highly customized advertising campaigns and doctors easily testing new drug alternatives, AI has opened many doors for businesses and individuals alike.
And the genetics field is no exception to this rule. AI can transform how we think about genetics and eliminate the boundaries that limited progress in research in the genomics area.
According to Science Daily, new studies already reveal that an artificial neural network has the ability to identify and uncover patterns in large amounts of genetic data, thus discovering groups and sequences of genes related to specific diseases.
Because diseases in an individual’s lifetime are “encoded” in their genetic sequence, the ability to understand humans’ genetic information on a highly granular level has come to be viewed as the key to generating treatments and solutions to such illnesses. The achievements in this area have historically been stunted by the fact that genetic data was too vast and complex for human evaluation; however, with AI’s ability to sort through and analyze a wide array of data in a short period of time — something that is nearly impossible for the human brain to accomplish as quickly — it has become possible to gauge into the genetic composition of individual people and achieve new heights in the area of genomics.
With AI’s ability to recognize patterns and learn on vast amounts of raw data, scientists have been mapping out biological networks based on the various interactions between proteins and genes: through deep machine learning, scientists have been looking into the possibility of identifying gene networks based on experimental data — something that hasn’t been widely applied in biological and genetic research in contrast to other areas such as, for instance, market research.
At the basis of this AI-driven research was a large database of genetic information — 20,000 genes, to be exact — from different people, the data being unstructured and reflecting the true expression patterns of the genes. With the algorithm lacking initial distinction between the genes from healthy people or people with various diseases, the AI network gradually learned to identify these patterns of gene expression encoded in the information.
Essentially, AI is on the way of allowing medical professionals to identify the exact gene groups that account for specific diseases — and it’s an enormous step towards finding the root cause and then finding an appropriate solution.
With the potential to peek into — and interpret — the individual genetic composition of any individual, AI also can accelerate the development of personalized medicine and treatments. What this means is the possibility of interventions uniquely tailored to a specific person or group of people with a similar genetic composition. While initially, individualization on such a high level was almost unheard of due to the astronomical costs and resources necessary, machine learning algorithms make it possible to tackle this issue quickly and efficiently, at minimal costs. While the algorithm identifies the common patterns within an individual’s genetic data, the system is then able to make predictions as to how likely the person is to develop a certain condition based on this analysis.
Examples of such AI applications can already be found in real-life use around the world. According to Forbes, the Canadian start-up Deep Genomics is already using its AI platform to decode the human genome in order to develop the best drug therapy or treatment for individual patients based on their cellular DNA composition. The startup’s AI-driven technology analyzes mutations and deviations, cross-compares the data to the thousands of mutation examples it has previously analyzed, and makes predictions of how the mutation can potentially impact a patient’s health.
Imagine if such technology could be applied to conditions and diseases that take away numerous lives each year. From cancer to Alzheimer’s and Parkinson’s, AI solutions can make it possible to find treatments for diseases previously rendered incurable and lethal.
The last — and potentially most controversial — application of AI in the genomics area is gene editing — or, simply put, the “editing” and alteration of a person’s DNA on a cellular level. While the idea of “editing” a person’s genetic composition or DNA poses a series of ethical and moral questions, it is a fact that AI can introduce a whole new level of precision to gene editing.
One of the more widely known technologies testing the boundaries of altering a person’s DNA sequence, CRISPR, “cuts” the DNA sequence in the necessary spots in order to stitch the material back together. However, the way the genetic material is stitched back together after the “procedure” isn’t as precise as desired; in fact, according to The Scientist, scientists have long thought that without a template, the stitching process is done rather at random. With AI, it will be possible to make more accurate — and safe — predictions on how to stitch the “cut” gene together.
Down the road, AI will continue to have massive contributions to genomics and genetics — from personalized therapy and treatment to disease control and even prevention, it promises to unlock new doors than no one thought of years back. And soon, AI will help us understand ourselves better — on a cellular, genetic level.
