GestaltMatcher Database - A global reference for facial phenotypic variability in rare human diseases

Area of Science:

• Medical Genetics
• Computational Biology
• Bioinformatics

Background:

• Facial dysmorphism presents significant challenges in diagnosis due to high phenotypic variability.
• Next-Generation Phenotyping (NGP) tools struggle with diverse patient populations, limiting their clinical utility.
• Genetic ancestry impacts facial features, complicating the recognition of syndromic patterns.

Purpose of the Study:

• To systematically analyze the impact of genetic ancestry on facial dysmorphism.
• To establish a diverse reference dataset (GestaltMatcher Database - GMDB) for rare genetic disorders.
• To evaluate the effect of data diversity on NGP tool performance.

Main Methods:

• Collected 10,980 frontal facial images from 8,346 patients across 581 rare disorders globally.
• Established the GestaltMatcher Database (GMDB) with increased representation from Asian and African populations.
• Analyzed NGP performance using diverse training and testing datasets, including non-European patients.

Main Results:

• Incorporating non-European patients into NGP training significantly enhanced diagnostic accuracy (+11.29% top-5 accuracy).
• Performance improvements were achieved without compromising accuracy for European patient data.
• The GMDB, adhering to FAIR principles, serves as a vital resource for clinical diagnosis and NGP advancement.

Conclusions:

• Cross-ancestral phenotypic variability in facial dysmorphism confounds NGP tools.
• International collaboration and data diversity are crucial to overcome NGP limitations.
• The GMDB provides a foundational dataset for improving NGP accuracy and clinical application in rare genetic disorders.