According to a story from news-medical.net, a research team affiliated with Princeton University has been hard at work developing a unique tool that utilizes machine learning in order to help identify genes that can cause a rare disease to occur. During a demonstration of the new system the team were able to successfully identify four genes that were associated with neuroblastoma, a rare type of cancer that mostly affects children. These genes had never been discussed in medical research before.
The system is called Unveiling RNA Sample Annotation for Human Diseases, or URSAHD. The tool functions by comparing RNA based gene activity to patterns that are associated with hundreds of other diseases. The ability of URSAHD to compare a sample alongside hundreds of known samples is an unprecedented achievement that is critical to the system’s capability. This process is essential in distinguishing unique aspects of different diseases and allows for an approach to research that allows for new information to be revealed that would be almost impossible to find using standard techniques. Needless to say, this approach should have exceptional value when it comes to rare diseases.
Another important detail is the fact that URSAHD uses RNA as its basis for examination instead of DNA. This means that the system can not only analyze genetic mutations but can also map out potential problems that can appear later, even from genes that at first appear healthy. URSAHD is an example of a precision medical tool, as it can distinguish between very similar diseases on the microscopic level in a degree of detail that other methods cannot.
The system draws on a sample of 8,000 biopsies that include publicly available records of gene activity related to rare diseases. The types of diseases that URSAHD can successfully analyze includes rare cancers, metabolic disorders, and heart disease.
The URSAHD system represents a significant leap forward in researching rare genetically-based diseases and disorders. The description of the system was first published in the scientific journal Cell Systems and can be found here.