This study implicates this variant as being much more common in a group of patients with Cowden or Cowden-like syndrome (which are inherited in a dominant manner, generally attributed to mutations in the gene PTEN) that lacked PTEN mutations and had elevated MnSOD protein levels. In a screen for SDHB mutations in these 74 patients, they found this variant heterozygously in two patients, and report not seeing it in 700 control chromosomes (equivalent to 350 individuals) (white of Northern/Western European ancestry).
They report a p-value of p<0.001 for the combined presence of this variant (H50R) and two other variants (G12S and S163P) in their cases (total of 8) compared to controls. This however is combining data for multiple variants. Counting alleles for this variant in particular, their data reports: case+ = 2, case- = 146, control+ = 0, control- = 700. A two-tailed Fisher’s Exact test gives p=0.03.
However, this study’s reported control data are suspicious and highly divergent from other reports — the authers state “all 700 controls had the entire sequence of SDHB, SDHC, and SDHD sequenced and no variants identified”. Both exome variation server data and low coverage CEU data report SDHB, SDHC, and SDHD variants with allele frequencies that diverge from the claim made regarding the 700 controls with extremely high significance (and, conversely, show no significant difference from the cases reported here). It seems very likely that some error was made in the analysis of the controls in this paper; without this statistical data the hypothesis that these genes are involved in Cowden or Cowden-like syndromes is seriously weakened.
At the time it was published, the authors were aware of already existing reports for the three most common variants of allele frequencies ranging from 1-3%. They forward two explanations for their controls being different, both of which are quite weak. (1) The ethnic groups samples (Spanish, French Canadian) may be significantly different from that studied in their paper (whites of Western/Northern European ancestry). This is implausible because later allele frequency data from the CEU cohort (which perfectly matches their description: Western/Northern European ancestry) confirms the 1-3% allele frequencies and is significantly divergent from reported 700 controls. (2) The 700 controls excluded individuals with symptoms of the disease, while other frequency data was from unselected populations. This is implausible because it posits a high penetrance dominant phenotype effect (obvious enough to cause exclusion of carriers) — this is a surprising hypothesis for such a common variant. If this were true, we would expect such an effect to be reported by now; indeed, we would expect significantly more of these carriers to be represented in the cases in this paper — this is not the case, frequencies in cases reported in the paper are not significantly different from the allele frequencies of the population at large.
The paper reports some functional findings, but in vitro findings do not necessarily predict a phenotype and should be interpreted with caution. (1) These variants are common enough that case/control or familial findings should be readily found and reproduced, and with data consistent with general population allele frequency data reported elsewhere. (2) In cases of extremely rare variants functional data may be relied upon to establish a particular variant’s effect, but in such cases other variants for the same gene (with similar functional data) should have a pathogenic effect supported with strong statistical evidence.