Molecular Biology

20190901
Kevin M. Wright, Kristin A. Rand, Amir Kermany, Keith Noto, Don Curtis, Daniel Garrigan, Dmitri Slinkov, Ilya Dorfman, Julie M. Granka, Jake Byrnes, Natalie Myres, Catherine A. Ball and J. Graham Ruby
https://doi.org/10.1534/g3.119.400448

The challenge of correlating lifespan with genetics, is that “genotypes are generally gathered from living persons, whereas lifespan (total elapsed time between birth and death) is a property of deceased persons. Due to this challenge, current age has been used as a lifespan proxy trait in many human aging studies.” However, as decades go by and additional studies are done using age as a lifespan proxy, there will be confounding factors having to do with such things as the advancement of medicine or increases in pollution.

While the genomes of children is related to those of the parents, the loci related to lifespan will be shared fractionally, and the fractions may vary based on the loci. Thus, a very large sample size is required for statistically significant conclusions. Several previous parental-lifespan studies have been conducted using the UK Biobank. To improve on the sample size, the data from AncestryDNA was combined with that from the UK Biobank.

While it is difficult to measure the genetic contribution to variation in human lifespan, the estimation is that the heritability effect of lifespan is under 10 percent. One confounding factor is social inheritance, where people are born into social factors that can contribute to better health or lifespan. Another confounding factor is assortive mating, a form of sexual selection where individuals that have similar appearance tend to mate with each other more frequently than if the mating was random (birds of a feather flock together).

Different races have different life expectancies, not only because of genetics, but also the socioeconomic factors they are born into. Isolating and quantifying the contributions to lifespan from these types of counfounding variables is not possible with current methods.

Phenotypes such as lifespan are influenced by more than one gene. Looking individually at each gene that affects lifespan, there are gene variants. Some of these variants result in the same phenotype, while some can cause differences in lifespan. The ones that do not cause differences are often grouped together, or classified as the same gene type, or allele.

Variants where a single nucleotide in the DNA is altered often result in the same allele. These single nucleotide differences are referred to as SNP (single nucleotide polymorphism) variants. SNP's can arise in any of the cells of the body due to mutations. While mutations are infrequent, they occur because the DNA copying process is imperfect.

As a population grows older, members of the same age with alleles that reduce lifespan, will die off first. The cells within an individual also mutate, but at a very slow rate. So slow, that DNA samples from the same person, taken 80 years apart, would still show a majority of cells having the original DNA.

There may be some cases, however, where mutations have a survival advantage compared to neighboring cells, and are able to gain majority. An example of this is cancer. However, it is unlikely that DNA samples will be composed of a majority of mutated cells.

Of interest is that loci associated with maternal lifespan do not correspond with paternal lifespan. “In total, this meta-analysis identified eleven paternal and four maternal lifespan-associated loci, two of which have not previously been associated with parental lifespan.”

Customer generated pedigrees are stitched together and unified, creating a duplicate-free ancestry. Recorded lifespans between 40 and 120 years were used, to reduce noise from deaths unrelated to old age. Approximately 1/3 of individuals before 1900 are missing a death date.

For example, some genes increased the chance Alzheimers while others increased the chances of cardiovascular-related deaths. One locus was linked to smoking behavior.

The APOE gene increased survival rate during mid-life, and yet increased the chances of death in later life.

However, 3 loci had no known disease associations. Namely the genes WAPL, SRRM3, and IP6K1 are associated with longevity but not with life-shortening diseases.

The increased number of persons from combining the two databases, led to a few different conclusions than previous studies on the individual databases. It could be that the sample sizes of the individual databases appear to be insufficient. In other words, the statistical power of the initial conclusions were too low. Or maybe certain genes could only possibly be of benefit in a British environment.

Another factor that could have led to the different results, is that the UK study used both the attained age of living and deceased parents, while the AncestryDNA study only had access to full lifespans.