Inborn Errors of Metabolism Knowledgebase
The IEMBASE accepts an array of biochemical and clinical symptoms from a user and returns a ranked list of possible IEM disorders that match the input profile. In addition, the system can explain the rationale of its results, suggest possible tests that would assist in narrowing down the differential diagnosis, and provide access to its database of biochemical, molecular, and clinical information if more evidence is desired.
Disease-characterizing profiles of clinical (n=1179) and biochemical (n=902) symptoms from an expert-generated database of 500+ IEM disorders were used. These profiles were mapped to the Human Phenotype Ontology (HPO) and Logical Observation Identifiers Names and Codes (LOINC) in order to exploit the semantic relationships of symptoms from the profiles. This, in turn, allows the expert system to determine a tiered list of possible IEMs which match the user-provided symptoms.