Lekárska genetika a diagnostika 1/2026
Automation of clinical interpretation of copy number variants
Copy number variants (CNVs) represent a significant class of structural variants associated with a broad spectrum of genetic disorders. Their clinical interpretation is a complex and time-consuming process requiring the integration of data from multiple database, functional, and population-based sources. Although the ACMG/ClinGen recommendations have introduced a standardized classification framework, a high proportion of variants of uncertain significance (VUS) and the demands of manual evaluation remain major challenges in clinical practice. In this review, we summarize key database resources and multilayer annotation approaches used in CNV interpretation and discuss their transformation into quantifiable attributes suitable for automated processing. Particular attention is given to hybrid frameworks that combine rule-based ACMG evaluation with machine learning–based predictive models. Such approaches enhance reproducibility, reduce the proportion of VUS, and preserve biological interpretability of results. We illustrate this principle with the example of the MarCNV tool, which implements automated scoring according to ACMG criteria, and the ISV model, which uses machine learning to predict the clinical impact of structural variants based on genomic attributes. The automation of CNV interpretation therefore represents a data-driven extension of clinical decision-making and a promising step toward precision genomic medicine.
Keywords: CNV, structural variants, ACMG classification, machine learning, genome annotation
