The role of phenotype-based search approaches using public online databases in diagnostics of Mendelian disorders

Purpose To investigate the effectiveness of phenotype-based search approaches using publicly available online databases. Methods We included consecutively solved cases from our exome database. For each case, the combination of Human Phenotype Ontology terms reported by the referring clinician was used to perform a search in three commonly used databases: OMIM (first 300 results), Phenolyzer (first 300 results), and Mendelian (all 100 results). Results One hundred cases were included (43 females; mean age: 10 years). The actual molecular diagnosis identified through exome sequencing was not included in the search results of any of the queried databases in 33% of cases. In 85% of cases it was not found within the top five search results. When included, its median rank was 61 (range: 1–295), 21 (1–270), and 29 (1–92) in OMIM, Phenolyzer and Mendelian, respectively. Conclusion This study demonstrates that, in most cases, phenotype-based search approaches using public online databases is ineffective in providing a probable diagnosis for Mendelian conditions. Genotype-first approach through molecular-guided diagnostics with backward phenotyping may be a more appropriate approach for these disorders, unless a specific diagnosis is considered a priori based on highly unique phenotypic features or a specific facial gestalt.