On 2015 Jun 02, Thomas Perls, MD, MPH commented:

An important conclusion of this work is that the genetic basis of surviving to age 90 years (the 5th and 15th percentiles of survival for men and women belonging to the 1900 birth cohort) is weaker and different from the genetic basis of surviving to the top one percentile, which in turn is different for those surviving to the top 0.1 percentile of survival.

The authors of a meta-analysis of largely nonagenarians (Deelen J, 2014) indicate that they were unable to reproduce the associations of the 281 SNPs in Sebastiani P, 2012 and used their negative result to assert that the findings were false positive associations. However, what this analysis of sibling relative risk of extreme longevity (Sebastiani P, 2016) indicates is that one should not be surprised by the lack of consistent results between these studies of people surviving to markedly different percentiles of survival and therefore having substantially different degrees of statistical power to discover variants associated with extreme survival. Reinforcing this point, when studies are similar in terms of a rare percentile of survival, a large number of the associated SNPs are actually replicated (Sebastiani P, 2013).

On 2015 Jun 02, Thomas Perls, MD, MPH commented:

An important conclusion of this work is that the genetic basis of surviving to age 90 years (the 5th and 15th percentiles of survival for men and women belonging to the 1900 birth cohort) is weaker and different from the genetic basis of surviving to the top one percentile, which in turn is different for those surviving to the top 0.1 percentile of survival.

The authors of a meta-analysis of largely nonagenarians (Deelen J, 2014) indicate that they were unable to reproduce the associations of the 281 SNPs in Sebastiani P, 2012 and used their negative result to assert that the findings were false positive associations. However, what this analysis of sibling relative risk of extreme longevity (Sebastiani P, 2016) indicates is that one should not be surprised by the lack of consistent results between these studies of people surviving to markedly different percentiles of survival and therefore having substantially different degrees of statistical power to discover variants associated with extreme survival. Reinforcing this point, when studies are similar in terms of a rare percentile of survival, a large number of the associated SNPs are actually replicated (Sebastiani P, 2013).