Friday, June 30, 2023

AI tool detects health-threatening genetic variants in vast haystacks.

Scientists have developed a way to sift through a person's genetic blueprint to find disease-causing variants. An issue that has frustrated doctors is that each human has an average of 4 million variants, sections of genetic code where we are different. Kyle Farh is the vice president of AI at the biotechnology company Illumina, where he and his team created an algorithm called PrimateAI-3D, using the genetic blueprints of 233 different primate species. The scientist looks for areas where the sequence is the same from one primate to another, which is a sign that any change is a problem. The new AI algorithm has a selection 1000 times as large as the archive used in most hospitals, ClinVar. PrimateAI-3D shows the three-dimensional structure of proteins, which can distinguish which mutations are harmful. 

They have tested the new tool on a biomedical database of more than 450,000 people in the United Kingdom’s Biobank. They found 97% of the population carried a rare variant that has an effect on health. With this new tool, they are able to predict health factors from the genome. Genomes are codes made up of four different chemical bases that build up DNA. One genome comprises approximately 3,200,000,000 nucleotides of DNA. In my opinion, having an AI model for something so complicated makes a lot of sense, and I don’t understand why it wasn’t implemented before. Machines have been running code in computer systems and figuring out the coordinates that send astronauts to space for decades. I think using artificial intelligence in this fashion is a great way to utilize the resources that are taking over many aspects of our daily lives.

No comments:

Post a Comment