RE: Germline Mutations in Shelterin Complex Genes Are Associated With Familial Glioma

The title of the paper by Bainbridge et al. (1) is an assertion that is not supported by the evidence presented. A genetic mutation is (positively) associated with a disease when a carrier of the mutation is more likely to contract the disease than a noncarrier or (equivalently) when a sufferer of the disease is more likely to carry the mutation than a nonsufferer. The results of the Bainbridge et al. study do not indicate an association.

The authors conducted whole-exome sequencing on 90 individuals from 55 families with a history of glioma and selected variants that were shared among two or more individuals diagnosed with glioma within a family. Two variants meeting these criteria were found, both in the POT1 gene. The authors went on to sequence the POT1 gene from a further 264 glioma patients. The two variants identified in the initial sample were not found in any individuals in the second sample. However, one patient was found in the second sample with a third variant in the same gene.

Normally, a scientific study involves forming a provisional hypothesis based on some data and then examining further data to see whether or not they support the hypothesis. The finding that no individuals in the second sample carried the variants identified in the first sample represents lack of support for the hypothesis that these variants are associated with glioma.

If the hypothesis being advanced is not just about the two variants found in the first sample, but is a more general hypothesis that mutations in POT1 are associated with glioma, then the data also fail to support it. The occurrence of a POT1 mutation in one patient out of 264 in the confirmatory sample is no more than would be expected by chance given the background occurrence of approximately 1% as indicated by the 65 nonsynonymous POT1 mutations found in the 6200 individuals of the control sample.

The occurrence of POT1 mutations in glioma sufferers in two families of the initial sample of 55 families does not represent evidence for an association, because the POT1 gene was singled out (out of a not precisely specified, but potentially quite large, pool of candidate genes) on the basis of mutations in this gene being carried by multiple glioma sufferers within a family in this particular sample of families.

Hunting for genetic risk factors in cancers is a fast-expanding field, and findings of “associations” are sometimes published before applying statistical logic to the evidence. It is troubling that the authors suggest that individuals with a family history of glioma should undergo screening for mutations in POT1. In the absence of evidence that such mutations confer an elevated risk of contracting glioma, such screening could cause some expense and much anxiety without an accompanying benefit.

The author is a statistician not active in the field of cancer research. He has a personal interest in the question of familial glioma, following two cases of glioma in his family. The study in question was brought to his attention by a concerned relative.