• Alex Dalgleish

Hear Melinda Mills discuss the intersection of genetics and fertility

Centre Director Professor Melinda Mills was the guest on the latest episode of the Insight podcast, discussing the intersection of demography and genetics.

Professor Mills joined host Razib Khan, a writer in population genetics and consumer genetics on his podcast, which aims to answer the question: ‘where do we come from?’ The pair discussed Professor Mills’ genetic research and what it tells us about human fertility today.

Razib said: “Melinda has bought this conversation to the next level by synthesizing demography and genetics,” and he was keen to get her thoughts on this fast-changing field.

Have a listen to the podcast, or take a look below for some highlights of the conversation.

Why apply genetics to human fertility?

I was initially looking specifically at reproductive choice and fertility from a social science point of view. When you look at it in this way, it’s often related to cultural norms, like gender roles and the labour market. Think about how the likes of childcare availability can affect the number of children you might choose to have.

I studied this for a while, and then I started to work with some biologists. I’d go to these interdisciplinary conferences, and they’d make a joke: ‘I want to introduce you to Melinda; she thinks fertility is entirely socially determined’, and everyone would laugh! And you start to think, this is a really blinded way of looking at the world.

So I started working with people from all types of disciplines, and narrowed in the obvious area of biology and genetics. I specifically wanted to look at people postponing when they have children and how many children they have – I thought that there there must be a genetic or biological component to that behaviour.

Trends in birth rates

When we look at fertility, we look at the timing of births and the number of children. The number of children people have has been going down over time and if we look at the average age of when people have their first child from the early 1900s until now there’s a u-shape. People don’t often realise that the age of having children at the turn of the century was in the late 20s, they think there’s always been postponement. But people started having babies comparatively early in around the 1940s and 1950s, before moving towards their late 20s and early 30s in later birth cohorts. There’s also a rise in childlessness, with around 20-25% of people born in the late 1960s in some countries not having children.

What we noticed was that in genetic studies so far, there has been a lot of focus on infertility and disease, but not on these demographic or behavioural traits. Our work looks at reproductive behaviour, which is age at first birth and number of children. These are actually very clear numbers - people almost always remember their age when they had their first or second child and the number of children they have! Those are very easy to report, so they’re kind of the gold standard in demography. In 2016 we published a study in Nature Genetics isolating genetic loci related to reproductive behaviour.

‘Hidden heritability’ in country and birth cohort

If we think about gene-environment interaction, these traits [age when having your first child and total number of children] definitely have a biological component, and definitely, there’s a genetic component – but actually, it has a lot to do with choice, and to do with your circumstances. When we think about complex traits like fertility and educational attainment, we need to look at them not only as genetic, but also contextually dependent.

We published a study in which we asked whether heritability differs across countries and multiple birth cohorts. Think about yourself, if you were deciding to have a child, and the norms and choices that apply – and then think about your parents. I think about my mother and when she decided to have a child. There were a lot of strict social norms about when she would have a child and how many she would have, which controlled a lot of her own ability to choose. My mother had her children in rural Canada; I had mine in the Netherlands and Germany in different times, much later and with less social and normative controls.

So we thought to ourselves, as social scientists working in genetics: would there be different influences on people in different historical periods and different countries? We also wanted to tackle the issue of hidden or missing heritability [in certain studies, heritability of these traits drops to as low as 1%]. What the genetic discovery studies often do is combine data from multiple countries and birth cohorts and assume that they’re not heterogonous – they don’t differ in any way.

In a study published in Nature Human Behavior, we looked across countries and cohort , and what we found was really interesting. The heterogeneity was what we thought – it was related people coming from different cohorts and national contexts, but it was really only for complex behavioral traits. For height, you couldn’t explain much from the country or birth cohort, but when you looked at educational attainment or age of child at first birth, around 40% could be explained by this ‘hidden heritability’ of the country or cohort you came from. By combining these cohorts and countries, we showed also through a series of simulations that you’re masking these differences in heritability.

Upcoming studies

We extended our previous 2016 study with two additional studies. One looks at number of children, linking it to ancient genome data and natural selection and the other examines age at first sexual intercourse and first birth. We found a strong genetic overlap in the genetic loci we found for reproductive behavior – they’re very much related to reproductive biology such as age of menarche, voice breaking for boys, but also externalizing behaviour like ADHD or smoking. But some things that we found (also highlighted in our 2016 study, and go into them further in a forthcoming study), is that the correlation between the loci that we found for age at first birth, have a 0.73 correlation with the educational attainment loci. It’s a scale of 0 to 1, so a 0.73 overlap is a very high genetic overlap between these traits. So in our forthcoming study we try to disentangle this. Did the educational attainment people actually find the fertility genetic loci? We’re also exploring things like age of onset of smoking and risky behavior – so there are a few surprises, and some things we will continue to work on. It is always work in progress.

Will the population keep growing?

People think that individuals will just increasingly continue to have children. But generally, people adapt to the environment. They want their children to go to school, or they want a nice car – and that competes with having children. We used to need children to care for us in

our old age, to help around the farm, or had many children because of high rates of mortality. When that falls away, what competes with children is often people’s own life goals or the fact that they think ‘one or two children is enough’.

It’s really interesting to predict the future, how this is going to work with fertility, and whether we reach these limits.

If we look at the estimates between the United Nations and others such as IIASA, there’s about a 1.2 billion person disparity between the estimates! The main reason is that people are have different estimates about will happen in Africa. Will girls get higher levels of education? Is this the key to overal lower levels of fertility? These are some of the really interesting and broader questions. I think the whole field needs to be a lot more global, particularly genetics. That is why we critically looked at this in some of our other work.

Future of demography

The geneticists call me a demographer, and the demographers now call me a geneticist! I think science is moving towards people who are multidisciplinary – or maybe going back to it.

That’s why I set up the Leverhulme Centre for Demographic Science, because it throws everything together. Sociologists, historians, geographers, zoologists, geneticists, statisticians. We’re all in it together, because that’s where the exciting discoveries come from, and that’s where we’ll break new ground.

The incentive systems within universities and journals aren’t quite catching up yet, we’re still judged in a very disciplinary way. So it has been an interesting and very difficult road, but I think it’s one that more of us will eventually follow.

You can hear the full podcast online now.


42-43 Park End Street

Oxford OX1 1JD

  • _github-512_edited
  • White Twitter Icon

© 2020 by LCDS