To the Editor:

In regard to the Viewpoint by Pantou et al. “The potential presence of the highly similar paralogue gene KCNE1B blurs the genetic basis of KCNE1-LQTS patients” [1], we have collected the following evidence in our laboratory, strongly supporting the possibility that the pseudogene KCNE1B, absent from the GRCh37 version of the human genome, is an artifact introduced in GRCh38:

  • We have evaluated the genotype of the rs1805127 SNP on 58 deaf patients analyzed with the OTOgenics Next Generation Sequencing (NGS) capture panel [2]. This SNP was selected because the reference nucleotide for KCNE1 (T) is different from the reference nucleotide for the paralogue position of the KCNE1B pseudogene (C), according to GRCh38 (Supplementary Fig. 1). Considering together all reads potentially originating from KCNE1 or KCNE1B, all patients show variant reads/total reads ratios of 0, 0.5 or 1, as it would be expected for the existence of a single locus with two alleles, instead of two paralogue loci with four alleles. If KCNE1B was not an artifact, since the sequences surrounding rs1805127 are identical between KCNE1 and the corresponding paralogue KCNE1B region, reads originating from all four alleles would be mapped together and KCNE1 heterozygous patients with a ratio of 0.25 should be identified (Supplementary Table 1).

  • We have evaluated a daughter-father-mother trio in which the father and the daughter are heterozygous for the rs74315445 KCNE1 SNV, while the mother is wild type. PCR was performed using primers that, according to GRCh38, should not be able to distinguish between KCNE1 and KCNE1B (Supplementary Fig. 1). Sanger sequencing of the obtained PCR products shows heterozygous peaks with identical heights for both alleles from the father’s and the daughter’s DNAs, which is not compatible with a variant affecting only one of four alleles. Although the father could be homozygous for the variant in the KCNE1 locus (which could explain the similar height of both peaks in his sample if KCNE1B was real), the mother is wild type and, therefore, the daughter could not possibly inherit two variant alleles from her parents (Supplementary Fig. 2).

  • We have PCR amplified the regions containing the positions for rs1805127 and rs1805128 SNPs on DNA from lymphoblastoid cell lines from ten individuals representing ten different ethnic origins from four continents in the same conditions unable to distinguish KCNE1 from KCNE1B. In all cases, Sanger sequencing of the PCR product shows clear wild type or heterozygous genotypes for rs1805127 (with peaks of identical heights for the alternative and reference KCNE1 alleles in heterozygotes) and homozygous KCNE1 reference (C/C) genotypes for rs1805128, with no sign of the reference KCNE1B (T) allele (Supplementary Table 2 and Supplementary Fig. 3).

In our opinion, these data, together with those presented by Pantou et al., strongly suggest that KCNE1B is an artifact that should be reviewed and curated. This would reduce the probabilities of false negative results caused by misalignment of variant reads originating from the true and clinically relevant KCNE1 locus to the KCNE1B region of the GRCh38 reference genome.