About the Institution

Germans Trias i Pujol University Hospital, located on the scientific campus near Barcelona, Spain, is home to a clinical genetics department that covers the full diagnostic journey - from sample preparation and sequencing through to reporting and genetic counselling. The department serves as the Spanish reference centre for phakomatoses, and leads ClinGen NF-SWN VCEP group.

The Challenge

A 12-year-old girl was referred to the genetics department presenting with macrosomia, macrocephaly, tall stature, overweight, learning difficulties, behavioral disorder, bifid-uvula, hamartoma, and abdominal hernia - a complex and ambiguous clinical picture.

The team followed their standard exome sequencing workflow: applying HPO-based filters, reviewing candidate gene lists, checking ClinVar variants, and manually curating results. Despite running through multiple layers of analysis - including mosaic variant detection - no variant emerged that convincingly explained the patient's phenotype. Standard approaches had reached their limit.

The Solution

With conventional HPO-driven and candidate gene approaches exhausted, the team turned to eVai's Suggested Diagnosis feature - an AI-driven layer that operates independently of predefined phenotype filters, scanning the full variant landscape for clinically significant findings that might otherwise go undetected, by taking into account pathogenicity, phenotypic similarity and inheritance pattern.

The result was immediate. eVai identified a variant in the androgen receptor gene - classified as likely pathogenic - that had not surfaced through any of the prior analysis steps.

The team manually reviewed and confirmed the classification, verifying domain rarity, PM2 criteria, and ClinVar evidence. The variant was confirmed likely pathogenic.

The Diagnosis

The finding pointed to Androgen Insensitivity Syndrome - a condition in which XY individuals are unable to respond to androgen hormones, resulting in female external genitalia and a phenotype that can include the features the patient presented with. The initial quality control had flagged an unexpected Y chromosome signal, but the clinical picture had not suggested this direction.

Once the diagnosis was confirmed, the clinical picture resolved: the patient's height, weight, and body measurements were entirely consistent with male growth scales. The learning difficulties were consistent with the syndrome. The family, who was previously unaware, received the support and counselling they needed to understand and navigate the diagnosis.

Dr. Elisabeth Castellanos - Head of the Clinical Genomics Unit

In Their Own Words

"eVai allows us to include different parameters to filter data (HPOs, frequency, protein effect) showing and linking the information available on different databases for each variant and disease identified. Its module of AI (Suggest Diagnosis) improves the standard filtering options by proposing other relevant variants related to clinical data, increasing the diagnostic rate" - Dr. Elisabeth Castellanos, Head of the Clinical Genomics Unit

Key Takeaways

• A flexible, two-step interpretation strategy - HPO-driven first, AI-driven second - enabled the team to solve a case that conventional approaches could not
• eVai's Suggested Diagnosis feature identified the causative variant after standard candidate gene and mosaic analysis had returned no clinically significant findings
• The diagnosis had direct clinical and family consequences, underscoring the real-world impact of interpretation tools that go beyond the expected