How is GestaltMatcher different from other tools like Face2Gene?

GestaltMatcher differs from many other facial phenotyping tools in several important ways related to data transparency, technology, and purpose:

• Algorithm and performance:

  GestaltMatcher AI uses an updated algorithm developed in 2023 that is particularly suitable for the recognition of ultra-rare genetic disorders.

• Data transparency:

  GestaltMatcher Database (GMDB), which is used to train and benchmark the AI, is a curated, FAIR-compliant dataset with openly available metadata. This allows external researchers to test and compare algorithms using the same data. In contrast, the data used by some other tools is not accessible, making independent benchmarking difficult.

• Privacy and data handling:

  When you perform an analysis with GestaltMatcher AI, patient data is not stored long-term. Our approach to data privacy and security is clearly defined. For other tools, data policies may not be as transparent.

• Explainability:

  GestaltMatcher incorporates ongoing efforts in explainable AI, including methods like Face2HPO, which aim to provide clearer insights into how the AI interprets facial features. This transparency supports clinical understanding and trust.

• Deployment options:

  GestaltMatcher offers an on-premise solution, which can be important for institutions with specific data security requirements. This solution has been described in the context of variant prioritization in VarFish in Bhasin et al. (2024; PMID: 38540429)

• Mission and governance:

  GMDB is an academic, non-profit project governed by the Association for Genome Diagnostics (AGD e.V.). This means it is focused on scientific collaboration and community-driven development rather than commercial interests.

These features reflect GestaltMatcher’s goal to provide a transparent, research-oriented platform that supports clinicians and scientists in rare disease diagnosis and discovery.