Domain Specific Language and variables for systematic approach to genetic variant curation and interpretation

Clinicians analyzing genomic data face three critical questions:  Is the variant present in the patient?  Does it affect the phenotype?  What evidence links the variant to the phenotype?  Answering these involves examining annotations encompassing provenance, confidence, and relevant biological and clinical evidence. To streamline this process, we have developed an Annotation Store —   a system that categorizes annotations by genomic scale, resolution, knowledge domain, and acquisition method, making them more accessible and actionable.  Building on the Forome AnFiSA platform (2019), we have enhanced the Domain-Specific Language (DSL) utilized by AnFiSA, integrating systematic annotations into curation workflows. The DSL represents rules as sequences of inclusion and exclusion criteria, facilitating the creation of lineage graphs that visually map the evidence for candidate variants.  The use of modern generative AI to write DSL scripts offers a transparent alternative to opaque AI models in genomic analysis. Unlike traditional models, DSL scripts are easily reviewable and approvable, making them ideal for peer and regulatory scrutiny while benefiting from the sophistication of AI-generated code.  By integrating systematic annotations with DSLs, we propose a forward-thinking solution for genomic workflows.