Computer-Aided Facial Analysis and Non-Coding Variants Detectable by Exome Sequencing Increase Its Diagnostic Yield: Results From the DECIPHERD Study.

Abstract

While exome sequencing (ES) reanalysis is recommended for patients undiagnosed after initial analysis, the combination of computer-aided facial phenotyping with targeted examination of non-coding variants remains underexplored. We aimed to identify the molecular etiology in undiagnosed cases from the DECIPHERD study (n = 167) through trio ES reanalysis guided by a computer-aided facial analysis. We selected 47 patients who remained undiagnosed after initial testing and whose facial photographs were available. Genes were prioritized based on facial analysis using Face2Gene. Subsequently, we scrutinized these prioritized genes for non-coding variants detectable by ES. This strategy increased the diagnostic yield by 0.6% (1/167). We confirmed the diagnosis of Kabuki syndrome in one patient harboring a likely pathogenic variant in KMT2D. Additionally, we identified compound heterozygous variants of uncertain significance in CEP290 in a second patient, suggestive of Meckel syndrome. As these variants remain functionally unconfirmed, this second case is reported as a candidate molecular diagnosis rather than an established one. This approach may represent a low-cost first step to unveil additional diagnoses without recourse to more expensive genomic testing.