Next generation phenotyping with quantitative narration for DEGCAGS syndrome

Abstract

The diagnosis of rare Mendelian disorders usually relies upon the interpretation of prose and is complicated by a lack of objective, reproducible phenotypic data. To address this limitation, we developed a next generation phenotyping workflow to phenotypically characterize developmental delay with gastrointestinal, cardiovascular, genitourinary, and skeletal abnormalities (DEGCAGS). We identified 15 people affected with DEGCAGS, including one novel patient identified at our hospital and 14 patients previously reported in the literature. Human Phenotype Ontology (HPO) terms were extracted from the patient chart and literature review. The HPO terms were sorted by count according to HPO hierarchy of terms. Phenotypes that cosegregate were identified utilizing a co-occurrence matrix. A quantitative narrative illustrated by violin plots was created for our patient from phenotypic data per each day of hospital admission. A total of 252 unique HPO terms were extracted from the patient record and literature review. The highest count of systemically sorted and unsorted terms and the most commonly co-occurring terms were described. A violin plot of phenotype occurrences demonstrated a progression of phenotypes over time. NGP offers a quantitative approach to phenotyping to generate phenotypic data in an objective and reproducible manner akin to NGS.

Keywords: DEGCAGS; bioinformatics; natural language processing; next generation phenotyping; phenotyping; quantitative narration.