Monday, November 13, 2017

Aging is a Group-Selected Adaptation, by Josh Mitteldorf

Josh Mitteldorf's Aging is a Group-Selected Adaptation places its thesis right in the title. Mitteldorf makes a strong case that aging is under the control of evolutionary pressures, and that the selection pressures for it are based on the benefits to groups, since it's clear there's no evolutionary gain to the individual. The evidence that aging is under evolution's control boils down to a comparison of many lineages that have long lives and have evolutionary cousins that do not. This is straightforward and hard to refute. The question is why.

The book's answer is that lineages that don't limit fecundity overshoot the carrying capacity of whatever environment they inhabit. The consequences are frequent population crashes. The alternative that leads to the possibility of stable populations is some feedback cycle that limits reproduction, combined with some way to ensure that deaths occur at a consistent rate. If the genes are optimized for the longest feasible life, then most deaths will occur in times of stress (resource exhaustion, unusual weather, or other cataclysm). This would lead to a much higher chance of ongoing boom and bust, which is a recipe for inevitable extinction.

There are some great graphs in the book illustrating the huge variety in life histories across many species. This one shows survivorship as a function of mortality and fecundity. When mortality is a horizontal line, survival falls consistently from birth to death. (hydra, hermit crabs, et. al.) Some species show decreasing mortality over their lifespan (desert tortoise, white mangrove, redleaf oak, ...), others only a slight uptick near the end (mute swan, tundra vole, sparrowhawk, ...).

According to Mitteldorf, the outcome of many experiments with artificial life show that one of the most valuable features of a species that has to cohabit with predators and prey is the ability to react to changes in its own population so that they have more progeny when the population density is low, and more individuals die when population density is high. The classical reaction to arguments about group selection says that this requires genes to have some kind of foresight, but the paradigm here is that populations that don't discover a way to reinforce this kind of response to population variation will be much more likely to go through frequent bottlenecks. Each bottleneck is another opportunity to go extinct.

One of the key ideas is that in order to contribute to ecosystem stability, rather than only to individual fitness, the genes must find a mechanism that leads to variation in robustness among the population. If some are slower, some are more susceptible to famine or cold, etc., then when a periodic stressor arises, some of the individuals will die. The alternative, if the genes design for uniform robustness is that all survive except when the stressor is severe, and in that case, nearly all will die. Aging, according to this thesis is a mechanism that causes variation within the population, ensuring a steady rate of death, which evens out rapid rises and falls in population. The population can still expand relatively rapidly when a niche opens up, but when living in a stable location, there are forces mitigating against population swings.

For those thinking about how to extend lifespan, a plausible first reaction to the idea that aging is selected for is to conclude that this means that aging will be harder to defeat. I would argue that the opposite may be true. Mitteldorf makes a good case that many lineages have found ways to allow individuals to live to arbitrarily long ages, so the biological mechanisms can't be infeasible or energetically unaffordable. Evolution's lesson is that we should be aware of the consequences of unlimited population growth, but given the demographic transition affecting most advanced economies, we can reasonably be more worried about the dangers of dropping population levels than of too many people. In any case, the hazards for human populations happen slowly enough that we'd be able to react before populations grow to be dangerous.

Aubrey de Grey wrote a response to Mitteldorf, but it looks like it was to an early version of the argument. (The book is dated 2017, but de Grey's 'response' is from 2015.) It looks to me as if de Grey had the reaction I described just above, and thought it was important to refute Mitteldorf's claims. I don't think de Grey directly addresses the arguments in the book. It seems to me that the argument presented here doesn't rule out the possibility of using de Grey's (SENS) approach to engineering fixes for the causes of aging, and it also provides for the possibility of other approaches that would directly intervene in the body's signaling that encourages aging and early senescence. If it's right, it doesn't reduce the number of possible approaches, it adds to them.