When we first saw this paper published, we noted that one of the PCR primers was identical to that published in our primary paper; the other primer had single base deletion. The remaining sequence facilitated their novel method, but that failed to determine accurately the region of overlap between Long and Very Long 523 polyT repeats. This group neither used our method in the so-called "non-conformation," nor did they take advantage of the commercially available test (costing less than $20 per subject) to obtain accurate results, even for a methodological comparison.

There are no quality assurance data for their genetic calls or their gel filtration method that would demonstrate the ability to size and call the critical overlapping sizes between the L and VL sequences that are biologically expressed in Caucasians. This wobble results in each sample not being informative in making L or VL calls. Even with additional ApoE genotyping, there is inaccuracy as clearly observed in the figures of Chu et al. The definition of age of onset is also not defined clearly or...  Read more

When we first saw this paper published, we noted that one of the PCR primers was identical to that published in our primary paper; the other primer had single base deletion. The remaining sequence facilitated their novel method, but that failed to determine accurately the region of overlap between Long and Very Long 523 polyT repeats. This group neither used our method in the so-called "non-conformation," nor did they take advantage of the commercially available test (costing less than $20 per subject) to obtain accurate results, even for a methodological comparison.

There are no quality assurance data for their genetic calls or their gel filtration method that would demonstrate the ability to size and call the critical overlapping sizes between the L and VL sequences that are biologically expressed in Caucasians. This wobble results in each sample not being informative in making L or VL calls. Even with additional ApoE genotyping, there is inaccuracy as clearly observed in the figures of Chu et al. The definition of age of onset is also not defined clearly or consistently and, similar to the ADNI data that are also inconsistently defined and applied, not suitable for age of onset curves. Cruchaga et al. (Cruchaga et al., 2011) subsequently reported a lack of conformation using these same ADNI samples that had non-standardized age-of-onset observations. The lack of consistency in the definition used for age of onset is a huge problem in retrospective collections unless there are specified prospective observations. Surely confirming age-of-onset data requires not just a test that functions with quality assurance, but actual accurate age-of-onset information. Additional age-of-onset curves illustrating the genetic relationships between the 523 polyT repeats and the prior highly reproducible genetic associations of ApoE4, ApoE3, and ApoE2 containing curves, derived from prospective clinical ascertainment and retrospectively analyzed from hundreds of subjects, were presented at AAIC 2012. We believe that these data, and the analysis of methodological differences, explain the authors' lack of confirmation, even of the ApoE3/4 effects.

View all comments by Allen Roses

Told you so (see reference).

I missed from the discussion a mention of the stochastic nature of gene expression, and specifically of the age at onset, as an explanation of the incomplete "penetrance" of mutations with a late expression. Penetrance is a dirty word that explains nothing, as many carriers die from other causes before expression of the mutation.

References:
Bruni AC, Montesi MP, Salmon D, Gei G, Perre J, el Hachimi KH, Foncin JF. Alzheimer's disease: a model from the quantitative study of a large kindred. J Geriatr Psychiatry Neurol. 1992 Jul-Sep;5(3):126-31. Abstract

View all comments by Jean-François Foncin