The Human Phenotype Ontology (HPO)—a standardized vocabulary of phenotypic abnormalities associated with 7000+ diseases—is used by thousands of researchers, clinicians, informaticians and electronic health record systems around the world. Its detailed descriptions of clinical abnormalities and computable disease definitions have made HPO the de facto standard for deep phenotyping in the field of rare disease. The HPO’s interoperability with other ontologies has enabled it to be used to improve diagnostic accuracy by incorporating model organism data. It also plays a key role in the popular Exomiser tool, which identifies potential disease-causing variants from whole-exome or whole-genome sequencing data. Since the HPO was first introduced in 2008, its users have become both more numerous and more diverse. To meet these emerging needs, the project has added new content, language translations, mappings and computational tooling, as well as integrations with external community data. The HPO continues to collaborate with clinical adopters to improve specific areas of the ontology and extend standardized disease descriptions. The newly redesigned HPO website (www.human-phenotype-ontology.org) simplifies browsing terms and exploring clinical features, diseases, and human genes.

Purpose This report describes the NIH Undiagnosed Diseases Program (UDP), details the Program's application of genomic technology to establish diagnoses, and details the Program's success rate over its first two years. Methods Each accepted study participant was extensively phenotyped. A subset of participants and selected family members (29 patients and 78 unaffected family members) was subjected to an integrated set of genomic analyses including high-density SNP arrays and whole exome or genome analysis. Results Of 1191 medical records reviewed, 326 patients were accepted and 160 were admitted directly to the NIH Clinical Center on the UDP service. Of those, 47% were children, 55% were females, and 53% had neurological disorders. Diagnoses were reached on 39 participants (24%) on clinical, biochemical, pathological, or molecular grounds; 21 diagnoses involved rare or ultra-rare diseases. Three disorders were diagnosed based upon SNP array analysis and three others using WES and filtering of variants. Two new disorders were discovered. Analysis of the SNP-array study cohort revealed that large stretches of homozygosity were more common in affected participants relative to controls. Conclusions The NIH UDP addresses an unmet need, i.e., the diagnosis of patients with complex, multisystem disorders. It may serve as a model for the clinical application of emerging genomic technologies, and is providing insights into the characteristics of diseases that remain undiagnosed after extensive clinical workup.

BACKGROUND & AIMS Autosomal recessive polycystic kidney disease (ARPKD), the most common ciliopathy of childhood, is characterized by congenital hepatic fibrosis and progressive cystic degeneration of kidneys. We aimed to describe congenital hepatic fibrosis in patients with ARPKD, confirmed by detection of mutations in PKHD1. METHODS Patients with ARPKD and congenital hepatic fibrosis were evaluated at the National Institutes of Health from 2003 to 2009. We analyzed clinical, molecular, and imaging data from 73 patients (age, 1–56 years; average, 12.7 ± 13.1 years) with kidney and liver involvement (based on clinical, imaging, or biopsy analyses) and mutations in PKHD1. RESULTS Initial symptoms were liver related in 26% of patients, and others presented with kidney disease. One patient underwent liver and kidney transplantation, and 10 others received kidney transplants. Four presented with cholangitis and one with variceal bleeding. Sixty-nine percent of patients had enlarged left lobes on magnetic resonance imaging, 92% had increased liver echogenicity on ultrasonography, and 65% had splenomegaly. Splenomegaly started early in life; 60% of children younger than 5 years had enlarged spleens. Spleen volume had an inverse correlation with platelet count and prothrombin time but not with serum albumin level. Platelet count was the best predictor of spleen volume (area under the curve of 0.88905), and spleen length corrected for patient’s height correlated inversely with platelet count (R2 = 0.42, P < .0001). Spleen volume did not correlate with renal function or type of PKHD1 mutation. Twenty-two of 31 patients who underwent endoscopy were found to have varices. Five had variceal bleeding, and 2 had portosystemic shunts. Forty-percent had Caroli syndrome, and 30% had an isolated dilated common bile duct. CONCLUSIONS Platelet count is the best predictor of the severity of portal hypertension, which has early onset but is underdiagnosed in patients with ARPKD. Seventy percent of patients with ARPKD have biliary abnormalities. Kidney and liver disease are independent, and variability in severity is not explainable by type of PKHD1 mutation;

Disease gene discovery has been transformed by affordable sequencing of exomes and genomes. Identification of disease-causing mutations requires sifting through a large number of sequence variants. A subset of the variants are unlikely to be good candidates for disease causation based on one or more of the following criteria: (1) being located in genomic regions known to be highly polymorphic, (2) having characteristics suggesting assembly misalignment, and/or (3) being labeled as variants based on misleading reference genome information. We analyzed exome sequence data from 118 individuals in 29 families seen in the NIH Undiagnosed Diseases Program (UDP) to create lists of variants and genes with these characteristics. Specifically, we identified several groups of genes that are candidates for provisional exclusion during exome analysis; 23,389 positions with excess heterozygosity suggestive of alignment errors; and 1,009 positions in which the hg18 human genome reference sequence appeared to contain a minor allele. Exclusion of such variants, which we provide in supplemental lists, will likely enhance identification of disease-causing mutations using exome sequence data.

Purpose:Medical diagnosis and molecular or biochemical confirmation typically rely on the knowledge of the clinician. Although this is very difficult in extremely rare diseases, we hypothesized that the recording of patient phenotypes in Human Phenotype Ontology (HPO) terms and computationally ranking putative disease-associated sequence variants improves diagnosis, particularly for patients with atypical clinical profiles.Genet Med 18 6, 608–617.Methods:Using simulated exomes and the National Institutes of Health Undiagnosed Diseases Program (UDP) patient cohort and associated exome sequence, we tested our hypothesis using Exomiser. Exomiser ranks candidate variants based on patient phenotype similarity to (i) known disease–gene phenotypes, (ii) model organism phenotypes of candidate orthologs, and (iii) phenotypes of protein–protein association neighbors.Genet Med 18 6, 608–617.Results:Benchmarking showed Exomiser ranked the causal variant as the top hit in 97% of known disease–gene associations and ranked the correct seeded variant in up to 87% when detectable disease–gene associations were unavailable. Using UDP data, Exomiser ranked the causative variant(s) within the top 10 variants for 11 previously diagnosed variants and achieved a diagnosis for 4 of 23 cases undiagnosed by clinical evaluation.Genet Med 18 6, 608–617.Conclusion:Structured phenotyping of patients and computational analysis are effective adjuncts for diagnosing patients with genetic disorders.Genet Med 18 6, 608–617.