Launching genetic diversity to the stars

Popular Mechanics asks, “How Many People Does It Take to Colonize Another Star System?”. The basic problem is that a multigenerational star voyage requires the trekkers to mate and reproduce many times while maintaining a limited population size. Too few people, and the colonists will rapidly lose genetic diversity by genetic drift.

The article starts by noting the work of anthropologist John Moore on the question. Moore concluded that the social structure necessary to prevent inbreeding was essentially that of clans or extended tribes of hunter-gatherers – strong kin avoidance rules to prevent inbreeding and a population size of 150-300 people.

A new paper by Cameron Smith focuses instead on the worst case scenarios, concluding that the “safe” population size would be much higher:

Entire generations of people would be born, live, and die before the ship reached its destination. This brings up the question of how many people you need to send on a hypothetical interstellar mission to sustain sufficient genetic diversity. And a new study sets the bar much higher than Moore's 150 people.
According to Portland State University anthropologist Cameron Smith, any such starship would have to carry a minimum of 10,000 people to secure the success of the endeavor. And a starting population of 40,000 would be even better, in case a large percentage of the population died during during the journey.

A number as large as 40,000 people would enable the mission to approximate the effective population size of the entire human population of earth before 100,000 years ago or so. For reasons I’ve discussed many times (for example, “Cultural impedance, demographic growth, effective population size”), the effective population size of humans does not mean that the actual number of people in the ancestral human population was very small. With Pleistocene people, there were many processes that reduced the genetic diversity (and hence our estimates of effective population size) within a population of a relatively large actual population size – on the order of a few hundred thousands of people.

Forty thousand is pretty small, but on a random-mating voyage of a hundred generations should basically approximate the Wright-Fisher population model. Smith further examines scenarios in which “catastrophic” events may affect the mission, greatly reducing genetic variation (or eliminating it). In these scenarios, a population dispersed across multiple “ships” would create a buffer, but each of those units has its own small population size issues, arguing for a bigger mission.

I’ll take a deeper look at Smith’s upcoming paper after the AAPA meetings. These future scenarios really help us think about the limits that existed in past human populations, which were less constrained in some ways but more so in others. Moore’s approximations for a future “generation ship” mission incorporated social dynamics in ways that have clear parallels in the past (his ethnographic work focused on small village societies of Southeast U.S. native peoples). Smith’s simulations refer to a larger-scale aspect of genetic drift.

The interaction of these two factors does not easily reduce to equations, but creates the most interesting anthropological questions. How much social control is necessary to maintain the viability of a colonizing population, not only genetic viability but also cultural viability? What is the balance between shared goals and practical needs?

I question the general assumption that such a mission would “need” to maximize the genetic diversity of the colonists. In fact, many potential groups of interstellar colonists might prefer to reduce their genetic diversity.

Imagine a small group of people with the sufficient motivation to divorce themselves from humans on Earth, launch across interstellar space for thousands of years, forcing their descendants to live within a tiny habitat, with the expectation that their common offspring will colonize a new planet a hundred generations hence. The kind of internal discipline necessary to motivate such a scheme is more like a cult than an open society.

Cults enforce cooperation by means of social isolation,

By increasing the relatedness of the population, they could enhance the incentives for cooperative behavior. In effect, people boarding the voyage on Earth would be assured that their descendants would not merely be notional descendants but in fact strongly genetically similar to them. Such groups don’t want to board this ship with a random selection of humanity; they want to board with their cousins. That reduced level of genetic variation would generate a larger genetic payoff for each individual launching from Earth.

They are not going to create a microcosm of Earth’s genetic variation. They’re going to create a colony of clones.

UPDATE (2014-04-05): A number of Twitter commentators have suggested that you don’t need to have so many people if you have a store of frozen sperm and egg cells. In essence, you could create a human version of the Long Term Evolution Experiment run by Richard Lenski. By unfreezing the eggs and sperm of previous inhabitants – or unrelated eggs and sperm brought from Earth – the colonists could add whatever genetic variation is required, or “rewind” the colony to a previous gene pool.

That can be done with today’s technology. We may question whether freezing is really a viable strategy across 1000 years. As yet we only know that freezing works over 20 years or so, and we don’t have good statistics yet about whether germ cells or embryos frozen over longer time periods have any increased chance of mutations or other long-term effects. Still, the level of risk already faced by interstellar voyagers is likely much larger than a slight increase in risk from long-term freezing of germline tissue.

It would make perfect sense to have a large store of adaptive variants available to deal with whatever challenges the colonists face on their new world. Imagine that they settle a world with only two-thirds of Earth’s sea-level atmospheric pressure. An influx of frozen Tibetan sperm would bring in genes to adapt the colonists to their hypoxic world.

Of course, we may also consider that a starship 50 or 100 years from now will be leaving an Earth with vastly greater genetic engineering potential than we currently possess. Colonists after the ninety-eighth generation might vastly prefer a bit of genetic tinkering to their own gene pool, instead of unfreezing vastly different DNA from an Earthling stranger. In that sense, the colonists will not need either a large population or a giant frozen sperm vat. They can build as they go.

This brings us back to social dynamics. The colonists must maintain their motivation and ability to put the colonization plan into motion as they arrive at their destination. Death of the colony is not the only risk; their culture may slowly devolve until they are nothing but interstellar lotus-eaters. We don’t know how large a cultural group is necessary to maintain the necessary traditions over a thousand-year voyage.

That seems like an interesting problem.