Between Natural Selection and Randomness: a Nearly Neutral Path to Genetic Diversity

• policy
• micro
• women

For Lovelace day 2015: Professor Tomoko Ohta, the proposer of the 'nearly neutral' model of genetic drift. This important modification to earlier theories describes the effect of random changes in the genome on the evolutionary direction of a species. Her work has implications for the speed of evolution in small populations (such as those being created scarily rapidly by ecological damage, for example) and to estimate how long it takes species to diverge from each other. This year, she received the Crafoord prize, one of the world’s biggest scientific prizes, from the Royal Swedish Academy of Sciences this year for her life’s work.

A Roundabout Route

Growing up in the 1950s, Ohta was part of the first generation of Japanese women able to study in good universities. She liked maths, but pragmatically wanted something with career prospects. Having failed an entrance exam to medical school, she went to study agriculture and work in publishing. After getting a job in plant cytogenetics, she was able to obtain a Fulbright scholarship and study in the US, eventually pursuing a PhD in population genetics.

This photo is kindly provided by Professor Joe Felsenstein, another famous geneticist, from his Flickr account.

When she returned to Japan, she sought to work with Dr Motoo Kimura at the National Institute of Genetics as he was "the only theoretical population geneticist in Japan at the time". In Ohta's own words (from an interview in Current Biology, paywalled):

At first, he was skeptical to let me do research in his field, but he finally accepted me as a postdoctoral fellow. Kimura was a typical Japanese man of his time, who regarded women's scientific activities as insignificant. After two years or so, I had convinced him that I should continue to do research.

No comment, although I hear anecdotally that he did come to respect her abilities and envy her invention of ‘nearly neutral’ theory. They spent a large part of their working lives researching and publishing together, and she also held him in high regard.

A Bewildering Diversity

Once people were able to study DNA in the wild, following its discovery by Watson, Crick and Franklin, they were surprised to find far more polymorphisms (differences between genetic code of different organisms within the same population) than expected. This really confused everyone because if selection occurred, genetic mutations should be fixed if they are advantageous or removed if they are deleterious: sticking around and making things more complicated for geneticists wasn’t really on the table. Some people suggested this could be due to balancing selection, a phenomenon where the direction of selective pressure keeps shifting from negative to positive, generating lots of variation. But this seemed unlikely to be responsible for all the polymorphism that was seen.

A resolution was proposed by Kimura, although it was controversial at the time. He proposed the ‘neutral theory of molecular evolution’ suggesting that most changes the molecular level were caused by random mutations with no beneficial or adverse effects to the organism. Because these mutations had no impact, i.e. they were neutral, changes in the gene pool occurred because of random processes such as genetic drift. In contrast, beneficial or disadvantageous mutations are increased or removed from the gene pool as a result of natural selection.

This made it quite clear that random neutral effects could explain a lot of variation, but it still couldn’t explain why polymorphisms were seen in genes and proteins which clearly had a function.

Nearly Neutral

In 1973, writing in Nature, Ohta proposed a refinement to Kimura's 'neutral theory'. Her paper suggested that if a mutation has a slightly negative effect on the performance of a biological molecule, there are likely to be many different ways that another mutation will restore its function. In fact, a slight setback opens the way for multiple different outcomes that diversify the population. With a more varied population, it is likely that some individuals will outcompete others and pass on this genetic spread, and evolution will be driven onwards.

This contradicts both the straightforward assumption of Darwinism that something unhelpful would set back that individual, and the neutral theory of molecular evolution developed by her mentor Kimura.

Her theory implies that since problems open the way for improvements, these can be carried forward despite being detrimental, and can become 'fixed' in the gene pool if the population is small enough.1 Later, she refined this to allow for slightly advantageous mutations that also open the way for further changes. Her theories address mutations that are border cases, neither completely detrimental, helpful nor neutral.

An example

In a 2002 review, Professor Ohta identified examples of biomolecules that seem to be showing the hallmarks of her theory.

Included was a 1998 study by Rutherford and Linqquist (paywalled) that addressed the heat shock protein Hsp90. This is a 'chaperone' protein, one that helps other proteins (in this case, ones responsible for signalling) to fold into the correct shape. This is particularly necessary if the other protein is a bit broken because of random mutations like those focused on by neutral and nearly neutral theory, and means that those changes basically aren't a problem.

However, if Hsp90 itself gets damaged, these hitherto unnoticable problems cause extreme differences in the large-scale organism. Fruitflies with this problem have extra antenna, deformed eyes, no eyes, and small wings, among others. This demonstrated experimentally that a small biochemical change, while initially neutral, can have dramatic effects at an organism level in some rare circumstances. While most of these small winged, myopic flies are obviously not very fit for purpose, if just a few of them happen to be super-flies it would be a major leap forward for the species.

So What?

Like most of the women you'll hear about on Lovelace day, she didn't have the easiest time because of her gender, and still gets rather overshadowed by her supervisor. However, she was the first Japanese woman scientist to be named Foreign Member by the American Academy of Arts and Sciences, also receiving the Japan Academy Award and receiving a “Person of Cultural Merit” award from the Japanese emperor in 2002. The receipt of the Crafoord prize this year is also a great step in for her recognition. But the most important thing here is her work.

It connects the microscale of evolution and life — just molecules and proteins, bobbing around — with the wide, beautiful and diverse world of interconnected organisms we see around us. The problem of how chemistry gives rise to efficient and diverse life is a deep one, touching on philosophy as well as chemistry, biology and statistics. The conclusion of most people in the field is that both random effects and the deterministic effects of natural selection are needed to produce the flowering diversity we see, and so theories that bring them together are precious. While it is still not clear what is causing polymorphism in a lot of cases, Ohta’s work laid the foundation for this synthesis: in her words, "Selection pressure seems to be inseparable from the force of drift."

As well as sounding like a motivational poster ('problems are opportunities to improve'!) 'nearly neutral' theory is interesting because it seeks to find a balance between random genetic drift and strongly driven selection, between simply categorising everything into 'help', 'harm' , 'neither'. It's an attempt to capture the reality of complexity rather than falling into easier simplifications, and that's awesome.

Many thanks to my friend Dr Mark Ravinet, a genetics researcher at the National Institute of Genetics (Professor Ohta retired in 1996 but is still seen regularly at NIG’s Journal Club) for both suggesting her as a subject, and help with wording and understanding. More from Mark here.

1. This just follows the law of large numbers: the larger the group, the more likely you are to get a representative sample and the less likely it is to see fluke effects. If you were to randomly choose 10 children from a school it would be more likely that they'd all be girls than if you pulled out 100. (Assuming it's a not a single-sex school). Equally, you're less likely to roll all sixes if you roll 100 dice versus 10. Small samples = greater chance of skewed results. ←