A natural model of population heterogeneity fits observed mortality distributions, including their tail
A.D.N.J. de Grey
Department of Genetics,
University of Cambridge,
Downing Street,
Cambridge CB2 3EH,
UK
It has long been appreciated that the Gompertz-Makeham distribution,
which very accurately describes observed mortality distributions over
most of their extent, conspicuously fails to do so for the longest-
lived subset of typical populations. Specifically, the "tail" of
observed survival curves is much longer than that of a Gompertz-Makeham
curve fitted to the whole population. A simple candidate explanation
is based on heterogeneity: the longest-lived subset of a population may
be enriched in individuals that intrinsically age more slowly than the
bulk of that population. However, a number of investigators have
concluded that this interpretation, while economical, is inadequate:
that it can explain some, but by no means all, of the observed deviation
from Gompertzian behaviour. Here it is shown that that conclusion may
be premature. Previous work has modelled heterogeneity in ways that
are computationally tractable but biologically oversimplistic; using a
more computationally challenging but also biologically more realistic
approach, however, it is found that observed distributions can be
approximated very accurately by purely "Gompertz with heterogeneity"
curves. Hence, no case can presently be made that some individuals
cease to senesce after a certain age, as has been proposed.
Key words:
mortality, Gompertz, heterogeneity
Problems or questions regarding this site should be directed to
webmaster@americanaging.org