RESULTS
Mutation screening in EFHC1 in our cohort of GGE patients revealed 11 heterozygous missense variants. They were found in 44 of the 100 probands with JME (44%), and one of the variants was also identified in 1 of the 7 individuals with GTCS on awakening (14%). Table 1 shows all 11 variants, along with their predicted protein repercussions and the clinical features of the carriers. All these variants have been deposited and are publicly available at http://bipmed.iqm.unicamp.br/GGE.
We revised and adapted the ACMG/AMP guidelines to our framework. For the pathogenic classification, we eliminated the following four criteria:PM1 (variant in mutational hot spot or well-studied functional domain without benign variation) because no mutational hot spots have been reported for EFCH1 , and the variants found in patients with GGEs are scattered throughout the gene; PM3 (variant detected in trans with a pathogenic variant), which is exclusive for recessive disorders; PP2 (missense variant in gene with low rate of benign missense variants and pathogenic missenses common) because EFHC1 has a low Z score for missense variants (0.14 according to gnomAD), indicating that the deviation of observed counts is not far from the expected number, and thus the gene is tolerant to missense variants; and PP4 (patient’s phenotype or family history highly specific for the gene) because GGEs are genetically heterogeneous (Zifkin, Andermann, & Andermann, 2005). We then organized the remaining 12 suitable criteria for a comprehensive classification scheme (Figure 1). We used the corresponding evidence of pathogenicity to meet the ACMG/AMP rules for combining criteria to classify variants (Table 2).
For the benign classification, we considered two ACMG/AMP criteria not applicable to our framework: BS2 (variant observed in a healthy adult with full penetrance expected at an early age) because GGE-relatedEFHC1 variants do not have full penetrance (Suzuki et al., 2004); and BP1 (missense variant in a gene for which primarily truncating variants are known to cause disease) because truncating variants in EFHC1 is rarely associated with JME (Bailey et al., 2017). We used the remaining 10 criteria in another classification scheme (Figure 2), an action that fulfills the ACMG/AMP rules for combining criteria to classify genetic variants (Table 2).
We assessed the different criteria for each of the 11 EFHC1variants found in our cohort. From the literature search, we observed that two variants, namely c.1765G>A and c.1820A>G, had not been reported. Next, we assessed the allele frequency of the variants in databases of individuals from different populations and computed the values (Table S1). All 11 variants were identified in at least one database. Nine of the 11 variants (82%) presented allele frequencies higher than 1% in at least one subpopulation, and six of them (6/11, 54%) had allele frequencies higher than 5%. Furthermore, we calculated the OR and statistical significance (P values) of the association between theEFHC1 variants found in patients with GGE in our cohort versus in a group of 100 control subjects of Brazilian origin and in two population databases of Brazilian individuals (BIPMed: Secolin et al., 2019; and ABraOM: Naslavsky et al., 2017) using an unconditional exact test (Z-pooled, one-tailed). The allele frequencies and test results are shown in Table 3. Moreover, we used 13 computer algorithms to estimate the deleterious effects of the identified variants (Table S2).
Finally, we classified the 11 variants identified in our cohort according to the ACMG/AMP guidelines (Table S3). The variants c.662G>A and c.685T>C were the only ones classified as ‘pathogenic.’ However, c.662G>A also meets the criteria for ‘benign.’ Six of the 11 variants were classified as ‘benign,’ and the remaining variants are considered variants of uncertain significance (VUS).