SRD5A3-CDG: 3D structure modeling, clinical spectrum, and computer-based dysmorphic facial recognition

Area of Science:

• Genetics and Molecular Biology
• Biochemistry
• Medical Genetics

Background:

• Steroid 5 alpha reductase type 3 (SRD5A3) pathogenic variants cause SRD5A3-congenital disorder of glycosylation (CDG), a rare inherited condition.
• While dysmorphic features are common in SRD5A3-CDG, a distinct facial recognition entity has not been established.
• Previous reports detail 43 affected individuals, highlighting the need for improved diagnostic tools.

Purpose of the Study:

• To report a novel SRD5A3 pathogenic variant and analyze its structural impact.
• To identify common clinical and dysmorphic features associated with SRD5A3-CDG.
• To develop and validate a computer-based tool for accurate facial recognition and diagnosis of SRD5A3-CDG.

Main Methods:

• Identification and characterization of a novel SRD5A3 missense variant (c.460T>C, p.(Ser154Pro)).
• 3D structural modeling of the SRD5A3 protein to predict the variant's effect on catalytic efficiency.
• Phenotypic analysis of patient data and published cases, including facial 2D image analysis.
• Development and validation of a computer-based dysmorphic facial analysis tool.

Main Results:

• A novel pathogenic missense variant, c.460T>C p.(Ser154Pro), in SRD5A3 was identified and structurally modeled.
• The p.(Ser154Pro) variant is predicted to be in a potential active site, potentially reducing catalytic efficiency.
• Common dysmorphic features include arched eyebrows, wide eyes, a shallow nasal bridge, short nose, and a large mouth.
• A computer-based facial analysis tool achieved 92.5% accuracy in recognizing SRD5A3-CDG.

Conclusions:

• The study integrates genotypic, structural, and phenotypic data for SRD5A3-CDG.
• A novel pathogenic SRD5A3 variant was characterized, providing insights into disease mechanisms.
• The developed computer tool significantly aids in the global diagnosis of SRD5A3-CDG by enabling accurate facial recognition.