July 30, 2019
Recently a young boy, Buddy, was suffering from a mystery illness at the Children’s Hospital of Alabama, and things were looking extremely serious. Young Buddy had previously been diagnosed as having a rare genetic mutation in the NGLY1 gene, but this current acute illness was unexplained.
Fortunately for Buddy his father was exactly the right person to help him. Matt Might is a computer programmer and head of the Hugh Kaul Precision Medicine Institute at the University of Alabama at Birmingham, and he uses AI to sort through medical information from millions of medical articles. The AI program, mediKanren, is still experimental, but Matt saw this is the only hope to save his son’s life.
He began to write code for mediKanren, to help sift through a huge array of biomedical information for possible clues for what was causing the acute illness. Buddy’s symptoms were entered into the system, and mediKanren began hunting for explanations. From the vast array of possible causes Matt homed in on causes of infection, as Buddy’s NGLY1 deletion means his immune system is compromised. The program identified a list of possible microbes that could be causing Buddy’s infection, including one called Pseudomonas. The medical team conducted tests, including metagenomic analysis, and found this bacterium in Buddy, as well as E. coli. When he was given antibiotics to target these two bacteria, Buddy improved rapidly. Without using AI the cause of infection may not have been discovered in time to save this young boy’s life.
There is an enormous amount of medical research out there, but without a way to sort through it all to find useful information it is not helpful. In challenging cases such as this one, using AI that acts as a reference librarian, sorting through information from databases across the world, interrogating millions of journal articles for medical information is an incredible step forward.
Source: STAT News
Read more: http://ow.ly/6NAG50vfqph
