CHOP Researchers Create Search Engine Linking Diseases and Observable Traits

Imagine this scenario: A patient comes to a doctor with a constellation of symptoms and physical traits that, at a first glance, do not lead to an obvious diagnosis. Perhaps the patient has facial abnormalities and cardiac symptoms, as well as other issues with the skeleton and joints. The provider suspects these symptoms relate to a disease that is due to an underlying genetic mutation but is unsure where to look.

A screenshot of the new search engine functionality Enter PhenCards, a new data resource and search engine created by researchers at Children’s Hospital of Philadelphia (CHOP) that links existing biomedical knowledge with observable human traits, also known as phenotypes. Pulling data from dozens of existing databases related to phenotypes, genetics, drug development, NIH funding and more, PhenCards allows researchers to home in on potential diagnoses, causal genetic pathways, and collaborators who are already studying the disease or ones like it.

“Providing all of this data in a single web server will allow clinicians, researchers and genetic counselors to add new layers of information to previously limited genetic studies and make more informed clinical decisions,” said Kai Wang, PhD, Associate Professor of Pathology and Laboratory Medicine at CHOP and senior author of the paper describing the new search engine. “We sincerely hope that with researcher and community involvement, we can add even more useful knowledge to our web server.”

PhenCards allows users to search by a single phenotypic trait – cleft palate, for example, or by pasting in the text of an entire de-identified clinical note, from which phenotypic terms are extracted by the search engine. The program’s disease prediction algorithm ranks the most likely related diseases and provides information about potential candidate genes and pathways, ongoing clinical trials, and potential collaborators and funders of future research. The site is especially useful for rare diseases, for which there may be limited information.

“PhenCards gives researchers the best possible chance of identifying a disease by comparing its phenotypic traits to similar diseases and even supplying novel candidate genes,” said first author James M. Havrilla, a bioinformatics scientist in Wang’s lab. “Our goal is to provide both a one-stop shop and a lasting, continuously updated resource that will further our understanding of human health and of both rare and common diseases.”

Learn more about PhenCards in the researchers’ paper, published in Genome Medicine.