Why we create Cardio-SM
The clinical interpretation of sequence variants in hypertrophic cardiomyopathy (HCM) is particularly challenging due to the marked genetic
heterogeneity of the condition, the low allelic recurrence of pathogenic variants, and the limited availability of robust functional and segregation
data. In the majority of cases, variants are observed only once (private variants), and population allele frequency data though informative are often
insufficient to support confident classification.
Furthermore, the sheer number of genetic variants present in each individual complicates the discrimination between pathogenic alleles and
benign background variation. This complexity necessitates a rigorous, standardized, and disease-specific framework for variant assessment, ideally
aligned with the latest ACMG/AMP recommendations as refined by expert curation panels.
To meet this need, we developed Cardio‑SM, a semi-automated variant classification tool that enables the structured application of ClinGen Cardio
VCEP specifications for eight sarcomeric HCM genes. The platform is intended to assist medical geneticists, laboratory professionals, and variant
curators in achieving consistent, reproducible, and evidence-based interpretations in cardiovascular genetics.
Variant interpretation
Variant interpretation begins with collecting basic information: the gene, the variant type (e.g., missense, frameshift), and its population frequency.
From there, you assess available clinical, functional, and computational evidence. Using structured frameworks like the ACMG/AMP guidelines and
disease-adapted tools such as Cardio SM helps organize this process into clear steps, ensuring consistency and accuracy in classification. The
implementation of Cardio‑SM began with the systematic mapping of ACMG/AMP rules to HCM-relevant gene contexts, using a structured,
multi-layered approach. We grouped the criteria into nine evidence domains, population frequency, prevalence, allelic origin, variant type,
functional impact, mutational hotspots, in silico prediction, segregation, and co-occurrence, each of which triggers associated rules based on the
cardiomyopathy-specific specifications.
An interactive user interface was then developed to guide curators through the classification process. The platform prompts users to input key
variant-level and case-level data or respond to conceptual queries aligned with the ACMG/AMP framework. Based on the selected evidence,
Cardio‑SM outputs a final classification—both under the standard categorical model and the quantitative point-based system—along with a
transparent summary of applied rules for downstream validation.
The tool, developed in Python and hosted on GitHub