Variants of unknown significance
Thirty-five unique VUS (with less than 2 LP/P interpretations in ClinVar) were encountered in our cases (Table S1). We aimed to perform RNA analysis in 3 of these, due to in silico prediction results (c.320-5T>A and c.1376-8T>C) or to the nature of the variant (duplication of exons 3 and 4). Lymphocytes for RNA analysis were available from one carrier of the duplication of exons 3 and 4, for several samples with c.320-5T>A and were unattainable from c.1376-8T>C carriers. RNA analysis showed that the duplication of exons 3 and 4 occurs in tandem and produces ∼30% of aberrant transcript containing an in-frame insertion of 273 bp (Figure 3). This affects the region that codifies for the FHA domain, unfortunately there were no polymorphisms in the region to perform quantitative analysis. This variant remains as VUS following all guidelines. Regarding c.320-5T>A variant, in silicoprograms predicted a reduction in the recognition of the splicing acceptor site of exon 3. cDNA analysis in two carriers showed the generation of an aberrant transcript, consisting in an in-frame deletion of exons 3 and 4 (Figure 4), as previously reported (Kraus et al., 2017). The amount of abnormal transcript seemed greater than 20%, although the absence of exonic polymorphisms prevented an accurate quantification. Of note, the frequency of c.320-5T>A is 0.12% in gnomAD (NFE) and of 1.35% (25 out of 1,848) in our HC cohort. In order to understand the differences in frequency in our population with relation to international databases, we screened 1,501 control samples. CHEK2 c.320-5T>A had a frequency of 0.8% (12 out of 1,501) in our controls, not a statistically significant difference, preventing it to be considered as risk allele.