Laura Wesseldijk works at the Max Planck Institute for Empirical Aesthetics in Frankfurt at the Behavioral Genetics unit in collaboration with the Department of Psychiatry at Amsterdam UMC. We talk about her research on the genetics of music and mental health, methods in behavioural genetics, the role of large samples, the importance of twins for behavioural genetics, and much more.
BJKS Podcast is a podcast about neuroscience, psychology, and anything vaguely related, hosted by Benjamin James Kuper-Smith.
Support the show: https://geni.us/bjks-patreon
Timestamps
0:00:00: Did Beethoven have bad genetics for music - or are there problems with applying (some) genetic methods to individuals?
0:11:51: Different methods in behavioural genetics
0:24:20: Gene x environment interactions and the difficulty of disentangling them
0:30:30: 23andMe in genetics research
0:37:26: Can you ask an interesting question if you need millions of people to have done a measurement?
0:42:08: How to measure musicality (at scale)
0:47:56: Geneticists really love twins
0:50:41: Do critical periods in music exist?
1:03:30: How Laura got interested in the genetics of music
1:12:07: A book or paper more people should read
1:16:17: Something Laura wishes she'd learnt sooner
1:17:49: Advice for PhD students/postdocs
Podcast links
Laura's links
Ben's links
References
Begg, ... & Krause (2023). Genomic analyses of hair from Ludwig van Beethoven. Current Biology.
Harden (2021). The genetic lottery: Why DNA matters for social equality.
Hjelmborg, ... & Kaprio, J. (2017). Lung cancer, genetic predisposition and smoking: the Nordic Twin Study of Cancer. Thorax.
Rutherford (2020). How to argue with a racist: History, science, race and reality.
Rutherford (2022). Control: the dark history and troubling present of eugenics.
Ullén, Mosing, Holm, Eriksson & Madison (2014). Psychometric properties and heritability of a new online test for musicality, the Swedish Musical Discrimination Test. Personality and Individual Differences.
Wesseldijk, Ullén & Mosing (2019). The effects of playing music on mental health outcomes. Scientific reports.
Wesseldijk, Mosing & Ullén (2021). Why is an early start of training related to musical skills in adulthood? A genetically informative study. Psychological Science.
Wesseldijk, Ullén & Mosing (2023). Music and genetics. Neuroscience & Biobehavioral Reviews.
Wesseldijk, Abdellaoui, Gordon, Ullén & Mosing (2022). Using a polygenic score in a family design to understand genetic influences on musicality. Scientific reports.
Wesseldijk, ... & Fisher (2024). Notes from Beethoven’s genome. Current Biology.
Laura Wesseldijk works at the Max Planck Institute for Empirical Aesthetics in Frankfurt at the Behavioral Genetics unit in collaboration with the Department of Psychiatry at Amsterdam UMC. We talk about her research on the genetics of music and mental health, methods in behavioural genetics, the role of large samples, the importance of twins for behavioural genetics, and much more.
BJKS Podcast is a podcast about neuroscience, psychology, and anything vaguely related, hosted by Benjamin James Kuper-Smith.
Support the show: https://geni.us/bjks-patreon
Timestamps
0:00:00: Did Beethoven have bad genetics for music - or are there problems with applying (some) genetic methods to individuals?
0:11:51: Different methods in behavioural genetics
0:24:20: Gene x environment interactions and the difficulty of disentangling them
0:30:30: 23andMe in genetics research
0:37:26: Can you ask an interesting question if you need millions of people to have done a measurement?
0:42:08: How to measure musicality (at scale)
0:47:56: Geneticists really love twins
0:50:41: Do critical periods in music exist?
1:03:30: How Laura got interested in the genetics of music
1:12:07: A book or paper more people should read
1:16:17: Something Laura wishes she'd learnt sooner
1:17:49: Advice for PhD students/postdocs
Podcast links
Laura's links
Ben's links
References
Begg, ... & Krause (2023). Genomic analyses of hair from Ludwig van Beethoven. Current Biology.
Harden (2021). The genetic lottery: Why DNA matters for social equality.
Hjelmborg, ... & Kaprio, J. (2017). Lung cancer, genetic predisposition and smoking: the Nordic Twin Study of Cancer. Thorax.
Rutherford (2020). How to argue with a racist: History, science, race and reality.
Rutherford (2022). Control: the dark history and troubling present of eugenics.
Ullén, Mosing, Holm, Eriksson & Madison (2014). Psychometric properties and heritability of a new online test for musicality, the Swedish Musical Discrimination Test. Personality and Individual Differences.
Wesseldijk, Ullén & Mosing (2019). The effects of playing music on mental health outcomes. Scientific reports.
Wesseldijk, Mosing & Ullén (2021). Why is an early start of training related to musical skills in adulthood? A genetically informative study. Psychological Science.
Wesseldijk, Ullén & Mosing (2023). Music and genetics. Neuroscience & Biobehavioral Reviews.
Wesseldijk, Abdellaoui, Gordon, Ullén & Mosing (2022). Using a polygenic score in a family design to understand genetic influences on musicality. Scientific reports.
Wesseldijk, ... & Fisher (2024). Notes from Beethoven’s genome. Current Biology.
[This is an automated transcript that contains any errors]
Benjamin James Kuper-Smith: [00:00:00] Yeah, I mean I thought we could actually start by, for me slightly unusually, just jumping straight into it and just starting straight with your, I guess one of your most recent at least papers, and presumably one that got you a fair amount of attention. Something that, I don't know, the headlines might be something like Beethoven not musical or something like that.
Not exactly that, but I guess it would be easy to construe it that way. So maybe do you want to just start by saying, What you did in that most recent paper, why and what you found.
Laura Wesseldijk: Yes. So, about two years ago now Begg and his co authors published a paper where they used hair strands of Ludwig van Beethoven and sequenced him to actually investigate some of his medical issues that he had. Whether he had a genetic predisposition for those because apparently just before Beethoven died, he wrote a letter to his brother asking him if they could [00:01:00] please find out what was medically wrong with him.
And so the authors of this paper investigated many of his medical issues, such as hearing loss and liver disease. But they didn't address Well, his trait, which he is most famous for namely musicality and his composing skills, of course and well, this wasn't very surprising because for analysis such as polygenic scores, where you calculate a person's genetic predisposition for diseases or behavior.
You need genome wide association studies, large genome wide association studies, and they're not yet available for musicality. But there is one G was available for whether people can clap in time to a beat, yes or no. So we thought I thought with a couple of colleagues, well, this is it.
Interesting. Let's see how [00:02:00] Beethoven scores yeah, his genetic predisposition to clapping in time to a beat. And I must immediately warn you that we, whether he was going to score low or high, we thought this actually to be. A nice illustration of what the pitfalls are of when you're applying these type of analysis, polygenic scores and results from GWASs on an individual level.
So we thought, let's use this as a, yeah, And valuable teaching moment to illustrate what we can and cannot do. And it turned out it was a really perfect example because Beethoven, we compared him, his genes to a sample of around 8, 000 Swedish individuals and around 6, 000 American individuals of which we knew , what levels of music they achieved in their life.
And Beethoven scored really low. He was in the ninth and the 11th [00:03:00] percentile in both samples. So you, I mean, it's practically, you can't really score any lower than Beethoven. So, yes, of course, this got a lot of media attention, and a lot of Beethoven does not have the genes and we were like, oh no.
So, but if you read the commentary, we really extensively discuss why this might at first seem really puzzling and surprising, but in reality isn't. So, I can tell you a bit more about the polygenic, there's, first of all, the limitations of polygenic scores. They are they do not capture the full genome.
They they make use of common genetic variants, and they do not include rare genetic variants. The
Benjamin James Kuper-Smith: the difference between common and rare?
Laura Wesseldijk: Well,
Benjamin James Kuper-Smith: I mean,
Laura Wesseldijk: It's in the name so the genetic variants that are common among the majority of people and rare genetic variants occur only [00:04:00] in in, yeah, a small group of people.
So, often in only maybe 1 to 5 percent of the population. And we know that these rare genetic variants often have larger effect sizes there but they occur way less often. So, GWAS studies only include common genetic variants, which occur in the majority of the population. And these genetic variances, variants have really small effects.
So, behavior is pleiotropic, which means that it is influenced by many genes, or many genetic variants, or with small effects. And also, all these genetic variants, they're not just influencing one. type of behavior, they influence many types of behavior. So it's really complex. So for G, GWASs, you need extremely large samples.
And as a result polygenic [00:05:00] scores are also dependent on these GWASs. So with the sample sizes of GWASs increasing, so will the predictive value of polygenic scores. And now I'm not talking about thousands of people. I'm talking more about hundreds of thousands of people or millions of people that need to be genotyped.
And you need to know in this case, where, how musical they are or yeah, other traits. So this of course is very difficult to achieve these type of databases. They're, they're work in progress. So that's A limitation of polygenic scores their predictive value is heavily dependent on the GWAS and the sample size of the GWAS.
Then they only reflect effects of common genetic variants and not rare genetic variants. Then of course, they're also dependent on the heritability of a trait. If you're looking at musicality we see, okay, it, it varies greatly what aspect of musicality you look at, but on average, there's a [00:06:00] heritability of around 40 5% while something like.
Diseases or height are way more heritable, so that would also, of course, influence predictive value of polygenic scores. Then there is an important thing that GWASs do not only reflect direct genetic effects, they also reflect interplay with cultures. So when you're applying polygenic scores based on a Western GWAS conducted culture, right now in time and you're applying it to someone who lived 200 years ago this might not be yeah, a good idea. And then besides these limitations of polygenic scores and why you shouldn't, I mean, they're okay. What are the most important thing, which I almost forgot to mention is that they're really population based approximations. So you can. use them in groups on a population [00:07:00] level, but if you apply them to individuals, they'll inevitably be individuals with a high genetic predisposition scoring low on the behavior and vice versa.
So you'll definitely it's, Yeah, it's not yet or maybe never a good idea to apply them on a person on an individual level. And then also in this case of Beethoven, we now used a GWAS, whether you can clap to a beat, yes or no, which of course isn't possible. going to be great at discriminating the higher levels of musicality and then also is composing music such as beethoven did really musicality or rhythm ability or is it actually more touching upon creativity for example so i mean there was also really The limitation of this G was, but overall, we thought it's a good yeah, learning moment to see that [00:08:00] this discrepancy between Beethoven one of the greatest musicians and composers in history and his mismatch with his genetic predisposition for something related to musicality, naming clapping to a beat that we should be very careful and very aware of the pitfalls of making individual level predictions.
Benjamin James Kuper-Smith: Yeah, I mean, so that's basically, yeah, I mean, that's why I, wanted to talk to you about, about this paper and obviously taking it more broadly. And I guess we'll, for the next hour or whatever it is, I guess we'll be going maybe more or less through many of the points you just mentioned in a bit more detail.
I had a kind of before we get properly into music and genetics I had a, just a brief question. So what does this Your example. What exactly does that tell you about the original? study that looked at Beethoven's health problems. I mean, for example, I mean, I didn't, I think I read the abstract.
I knew that one, I think, it was something like they found a really high [00:09:00] predisposition for liver cirrhosis, which is not surprising when you consider that, well, for one, I don't know whether that's a coincidence, but Beethoven's father was a raging alcoholic. So, I don't know, it makes sense when you know Beethoven's life story.
But and Beethoven having had so many health problems in his life, I'm, yeah, I'm curious kind of does, to what extent do some of the limitations you just mentioned also apply to the original study in a slightly different context? Or
Laura Wesseldijk: Well, the original study had a way more genetic techniques to actually look at his DNA. They also focused more on monoallelic diseases and they did make use of, so this really applies only to polygenic scores. And to applying polygenic scores on an individual level, and the original paper was way more extensive and elaborated, elaborative than that.
So, um,
Benjamin James Kuper-Smith: as a kind of broader question to not make it maybe too specific to that other paper, but [00:10:00] when, under which circumstances would it be sensible to do something like what you did? And for the results to actually be meaningful and to tell you something interesting where you look at someone, an individual's genetic predisposition for something what you need for that to actually be sensible.
Laura Wesseldijk: Well, you need diseases disorders. I'm not sure if there actually is really human behavior because most human behavior is really complex and they're complex traits. But you would need, I think probably traits and diseases that yeah, are strongly influenced by genetics by, Hopefully a couple of genes only with large effect sizes.
For example, when you look at Alzheimer's disease or when, Huntington's disease, when you have carrier status and you have a certain version of an allele. And when you have those alleles your risk is really significantly increased. Like, [00:11:00] those type of diseases and traits. then it can be useful to look at genetic predispositions.
Of course, also, I think, for complex traits and human behavior. I mean, practically all human behaviors influenced by genetics. And I think polygenic scores are still very valuable to look then in larger groups and see how it interacts with the environment. But on an individual level to make predictions from birth onwards, how it's going to go, then of course, it's always an interplay between genes and the environment.
So I wouldn't apply it to, to any human behavior at this point in time.
Benjamin James Kuper-Smith: Okay. That's a good answer. That's easy to remember. Yeah, so maybe a little bit about the let's see which order. Um,
Yeah maybe as a kind of, to take a little bit broader first on genetics and the kind of studies you do. Or not just you specifically, but the field. Um, I mean, you [00:12:00] mentioned now the what's called genome, genome wide association studies what different techniques are there?
I mean, I remember, twin studies are always fun. I wonder actually one question I have there is how. how relevant and useful they are, or whether that's more like an older construct that's been phased out by the more modern techniques. Yeah, what are the kind of main ways in which you can study the genetic effects of something and maybe also what are the best ways of doing it today, or what does it depend on?
Laura Wesseldijk: Yeah, so, I mean, there are individual differences in human behavior such as musicality. Some persons are really great playing the piano while others are not at all good. And most people perform on an average level and this individual variation can be caused by environmental factors. Factors and by genetic factors and to investigate to what extent genetic factors are [00:13:00] of play the field of behavioral genetics can make use of twin data or of genotype data.
And twin data, twin studies have been around the longest already since the early 1900s. And they make use of the fact that there are two types of twins. So they're identical and non identical twins. And identical twins, they develop from the same fertilized egg and they share 100 percent of their DNA, while non identical twins, just as regular siblings both develop from their own egg and own sperm cell, and so they share on average 50 percent of their DNA.
But both type of twins share the family environment they grow up in. And they also both are exposed to unique environmental factors that are unique to them and not shared with their co twin. And because of this, because of the existence of identical and non identical twins we can actually estimate exactly to [00:14:00] what extent genes are at play on any type of human behavior, diseases, disorders, traits that you can measure in people.
So when the resemblance within identical twins is higher than within non identical twins, we infer that genetics play a role. And then when the resemblance is closer to each other, the family environment, the rearing environment is of play or plays at least a role to some extent. And everything in which identical twins differ from each other is due to the unique environmental influences.
And with a bit more complex statistical models you can actually exactly get an estimate of heritability of the influence of genetic factors and the family environmental factors and the unique environmental factors. So this is really used to estimate heritability. Now, [00:15:00] then after, I think, decades of twin studies showing that practically all human behavior is to some degree influenced by genetics.
The DNA helix was discovered in 1953.
Benjamin James Kuper-Smith: So just a brief question there, isn't I also vaguely remember adoption studies. Are those common or are they so, I mean, I remember like the holy grail of twin adoption studies or something like that, but I'm, is that just so rare that
Laura Wesseldijk: It's not
Benjamin James Kuper-Smith: not useful? Oh, it's not allowed?
Laura Wesseldijk: no, you're not allowed to separate twins from each other anymore. This has happened in the
Benjamin James Kuper-Smith: But it, it happened, right? Yes.
Laura Wesseldijk: It has happened indeed, but it's no, it's not it's not allowed. So we don't have current adoption studies. Based on twins separated from each other? No. So, but yeah, they do exist.
Benjamin James Kuper-Smith: so they did exist, but it's not really much of a fact okay, good.
Laura Wesseldijk: So then after the discovery of the DNA helix in 1953, the field of molecular genetics arose.
And [00:16:00] Then actually first came linkage studies that tried to link certain parts of the genome to behavior. And this turned out to be actually only successful in case of genetic variants with very large effect sizes. So monoallelic diseases, for example. So this is really an exception. This most human behavior doesn't work like that.
And then came actually candidate gene studies which we know nowadays also that suffer from replication issues because people made yeah, a priori hypothesis on this part, this genetic variance is. is associated with this behavior. Let's test it. By that time people didn't know enough yet about the genome to actually make really good a priori hypotheses.
Because nowadays we know that a behavior is pleiotropic. So influenced by many genetic variants with very small [00:17:00] effect sizes. So the candidate gene studies were really, Yeah, really failed and have replication issues, but then the first whole genome was sequenced and that allowed genotyping to actually become cheaper and faster because so our DNA consists of around 4 to 5 million single nucleotide polymorphisms, SNPs, And after more knowledge about the human genome SNPs could also be imputed.
So genotyping became yeah, more accessible. And then this actually gave rise to genome wide association studies. So GWASes. And what happens in these GWASes? is that groups are compared a group, for example, that can clap to a beat versus a group that cannot clap to a beat. And all the snips are compared for an association with this trait, whether you can [00:18:00] clap to a beat, yes or no.
So this leads to an enormous amount of tests. So there needs to be really rigid control for multiple testing. correction for multiple testing. And but these GWASs actually provide associations with genetic variants with behavior that are replicable and it's an hypothesis free design, but also it quite quickly became clear that you need extremely large samples because of these genetic variants having such small effects.
You really need, yeah, preferably millions of people genotyped. With information about the trade that you're interested in. So this is of course musicality wasn't highest on the priority list. This has mostly of course, first been done for diseases, disorders, psychiatric problems. Also human traits nowadays, height, very well investigated [00:19:00] trait.
But musicality Yeah, there's currently actually a musicality genomics consortium that's working on this. I'm also part of that. So we try to really collect data from all over the world with where there's data available, genotype data available of people that we also know musician status from. And then. After these GWASs, actually a lot of post GWASs approaches developed, such as the polygenic score approach, which makes use of the results that come from a GWAS. The effect sizes that are found in a GWAS are then actually applied to an individual person. So for anyone who you then have genotype data from, you can calculate his or her genetic predisposition for the trait that was studied in the GWAS.
So you compare the DNA of the participant, of the individual what genetic variants they have on what genetic loci [00:20:00] and what the effect size is in the GWAS. And then you add those up all together. And one score comes out representing the genetic predisposition for the trait in the GWAS.
And now. You would think that yeah, your question was, what do you use nowadays? Do we use still use twin studies or are they now overruled by molecular genetics? But actually there's. What is really great about the last probably 10, 20 years is that there has been a bit of a merge between twin and molecular genetics because you can use twin data and family data to actually further investigate GWASes.
You can use polygenic scores within families and compare them, for example, between non identical twins to control for, yeah, certain indirect genetic [00:21:00] effects, such as passive gene environment correlation, which can happen because genetics can also have an influence through environments on the child.
If you have parents that have genetic predispositions for music, but they also offer their child a musically enriched environment, then this actually inflates the effects of genetic variants as detected in GWASs. So with twin data, you can actually control for this. So nowadays there are a lot of within family GWASs done to actually try and control for these indirect genetic effects. And also what is really, I think, very important is that twin studies. When we think of twin studies, we think of heritability estimates, but they can do way more than just a heritability estimate, which in itself isn't that informative. Whether something is heritable, yes or no. Heritability estimates are also very dependent on the age [00:22:00] of your sample size, the country that you're actually conducting the study in.
They differ. So something can be heritable. Yes, but it isn't. As informative to know, oh, it's 46 percent heritable. But what twin studies can do more is they actually also allow the opportunity to investigate. causality, because identical twins are the perfect control for each other. They're like yeah, they're almost like a perfect randomized control trial.
Because if you compare monozygotic twins that differ in something, for example, one smokes and the other doesn't, and you really truly want to investigate whether smoking causes lung cancer. You have to see whether the association is still significant identical twins that differ on smoking because otherwise you might say it's the genes, it's overlapping genes or family factors that influence both smoking and lung cancer.
But no, [00:23:00] actually this design was used to show that smoking causes lung cancer. So, twin studies can be used for that. Twin data can also be used to disentangle gene environmental interplay, such as interactions between genes and the environment, when genetic influences vary dependent on an environment that you're exposed to or not or gene environment correlations such as Genetics actually actively creating certain environments around you.
When you have a genetic predisposition for musicality you possibly have way more CDs in your house and a piano, and you actively search for things to enrich your life musically. And things like that. twin studies can still easily be used for. So it isn't it isn't that one of the two fields has won.
It's more, they both have their advantages and disadvantages, [00:24:00] and they also really now merge together. which I think is a very nice development of the 21st century. We can combine molecular data with twin data and actually even better investigate certain gene environmental interplay.
Benjamin James Kuper-Smith: Okay, cool. Yeah, so there's still a use for twins. That's nice. Yeah, I mean, you already explained it here and there earlier, but maybe just to do it once. I think, In one of your articles you mentioned there are these three types of gene environment interaction. Do you want maybe with the example of music just go through the, what was it, active, reactive, passive, something like that.
Yeah, what are the three and how can genes and the environment interact?
Laura Wesseldijk: Yes. So, okay. So first of all, gene environment correlation is when genetic and environmental factors correlate. So they're associated with each other, and you have three types. There's a passive gene environment correlation, reactive gene environment correlation, and active gene environment correlation.
And passive gene [00:25:00] environment correlation refers to the genotype that you inherit and the environment in which you are raised that they are correlated. So for example, your parents have a genetic predisposition for musicality, but also provide you with a musically enriched environment. So you inherit both.
the genes for musicality, but you also grow up in this musically enriched environment. Reactive gene environment correlation is when the environmental experiences and conditions that are provided to a child are actually a reaction to the child's genotype. So for example, I have inherited a genetic predisposition for music and my peers and my family keeps providing me with music classes, for example, or a piano instead of a football, because I somehow seem to have an interest for music more so than for sports.
And active [00:26:00] gene environment correlation is when. There is a correlation between genes and the environment, certain environments. You'll see that genetic variants more often occur in certain environments because this genotype causes people to actively search for certain environments. So I would yeah, I go to concerts a lot and I hang out at these music schools and in cultural houses.
Because of my interest and talent for music. Well, instead, if I would have had more of a genetic predisposition for physical attributes that would make sports more attractive to me, I would yeah. Be at soccer games, for example. So, These are all examples of how genotypes and environments are related and correlated and intertwined and therefore also bias research findings because,
Benjamin James Kuper-Smith: Yeah.
Laura Wesseldijk: yeah, there is [00:27:00] an association with genetic factors.
Benjamin James Kuper-Smith: Yeah. I mean, it's, I mean, I know we you already mentioned a little bit earlier, but maybe we can go through it a little bit again, because it seems to me like such a difficult problem to solve, right. To disentangle these three, because I mean, when you basically went through the three examples, I mean, that was basically exactly my life in that sense, because I almost became a classical musician and, my parents both did music as a hobby.
So we had instruments here and then I picked it up through that. And then I progressed faster than other people. So I got more comments on that and blah, blah, blah. And it was clearly something worth from other people putting time and investments in. And then I tried to go to orchestras and well, that kind of exactly what you just mentioned.
Right. So like, I mean, do you need them the rare case where someone randomly has a genetic mutation so that their family is not interested at all? I mean, yeah, how do you basically disentangle these three? Because it seems to me that probably often very highly correlated.
Laura Wesseldijk: [00:28:00] Yes, indeed, they're very highly correlated, and they're difficult to disentangle. So one very interesting research method that has developed over the last five years is with the use of molecular genetic data and calculating polygenic scores. And then seeing whether they predict the outcome. So in this case, for example, polygenic score for musicality, whether it predicts musicality among independent, unrelated people, versus whether this polygenic score predicts musicality within family members.
In that case, you keep the passive gene environment correlation stable because these family members, these siblings or non identical twins are raised in the same family. Whether the predictive value decreases or increases gives you an yeah, suggests indication whether there is passive gene environment correlation.
However it is [00:29:00] very difficult to disentangle because it can also be actually confounded by things such as population stratification and assortative mating, which are concepts that actually also influence effect sizes off genetic variants. That's maybe a little bit too technical for now, but I think it's also I mean, we don't necessarily need to disentangle which type of gene environment correlation is going on exactly.
It is just very necessary to be aware of it when, for example, you're looking in populations and you're studying things and you're making conclusions about, um, certain effects I, yeah, I can't think of a great example right now, but about well, yeah, people in music schools if they live close, do they already have a, an something a, an a talent for music or when you're just, when you're analyzing things and doing [00:30:00] research, you just need to be aware that there can be a bias.
of this intertwined gene environment correlation. So what type is exactly at play is maybe not the most interesting question, but when you're trying to make important conclusions about effects of certain interventions, cultural interventions, or music interventions, it's important to take this into account.
Benjamin James Kuper-Smith: Okay. To pick up on something you mentioned earlier briefly, which is, The accessibility of genetic testing has increased. So I saw that on some of your papers, either 23andMe was a in the acknowledgements, or sometimes they have like a research team that was part of the co authors and that kind of stuff.
Yeah, I'm just curious. I always assumed that 23andMe was a Something like 23andMe was mainly like a marketing gimmick that somehow people managed to convince other people that this was actually useful. But maybe from a researcher's perspective, in terms of using this [00:31:00] data or working with it or whatever, like how, yeah, can you just describe like maybe how you started collaborating or working with them and what the, how they, yeah, I guess how they can help you do research better.
Laura Wesseldijk: So 23andMe is this genetic company, or it's a company that asks for your, well, you pay them to, and provide them with your DNA and they report back on your ancestry and on health status. Issues genetic, that your genetic predispositions for certain health and behavioral traits.
And this has been done by millions of people. So they have a lot of information because also some of these people filled in questionnaires, answering the question, whether you can clap to a beat, yes or no. So actually haven't worked together with 23andMe, but they actually [00:32:00] run GWASes, or let researchers run GWASes.
And they, you can apply to receive the summary statistics from those GWASes. and then use them in further studies to create polygenic scores. And that is what I have done. So then you need to obviously acknowledge them or either make them a coauthor or acknowledge them. But I haven't worked together with researchers at 23andMe.
But I have made use of their data and of their results from their genome wide association studies. I actually have only recently also sent my own DNA data
Benjamin James Kuper-Smith: I wanted to ask whether you've done that, yeah. Yeah.
Laura Wesseldijk: I did it around the same time the Beethoven paper came out. And in the Beethoven paper, I really tried to make a point of do not make individual level, [00:33:00] predictions do not yeah, predict, make genetic predictions for an individual.
And this is exactly what 23andMe does. I mean, they also have a carrier status information. So in this case, you can see whether you carry Certain variants for Alzheimer's Parkinson's that's different, but they do make use of polygenic scores as well. And I caught myself reading that my preferable time to wake up is five past eight.
And I set my alarm clock to five past eight, because obviously this is what my genes tell me. And no, of course, I mean, as I already explained before we can't make these type of predictions based on polygenic scores. And there will inevitably be a lot of wrong predictions. And, um, yeah, people scoring high and low on traits.
They clearly have. Yet I did find it really interesting.
Benjamin James Kuper-Smith: so presumably you have a higher predisposition to clapping to the beat than [00:34:00] Beethoven? Did you check that?
Laura Wesseldijk: They didn't report that they actually didn't but they do have they have information on a musical trait and I scored higher than average and I thought, see, I know, but obviously, this is, it's not, I shouldn't think like this,
Benjamin James Kuper-Smith: But why did you send it in then? Was it for the Some of the health stuff, or
Laura Wesseldijk: I got it as a present, first of all, but I also, I did, I was very curious actually mostly curious about ancestry results
Benjamin James Kuper-Smith: Okay.
Laura Wesseldijk: then the health things and the traits are just really fascinating to see, especially because I work with these data myself as well. And I'm very aware of most of the GWASs that they have used.
So I thought I was, it was just curiosity and a bit of fun, but obviously I haven't. Okay, except for my alarm clock made any major decisions for my daily life.
Benjamin James Kuper-Smith: So, random question about this. I vaguely have in memory that, for the ancestry stuff, that they It [00:35:00] depends on where the people Like, they compare it to the origin of the people who send them their data. So, I think I heard something once that, for example, in Germany we have lots of Turkish immigrants, historically.
So that sometimes you get, like, if you're completely German, You get a bit of Turkish because lots of people who have Turkish ancestors now live in Germany. So there's this weird, I don't know, which sounds like, a bit of an error in the correlation in terms of this kind of stuff.
I'm just,
Laura Wesseldijk: I don't know exactly about that. But I do know that this leads to a huge under representation of certain countries. So for example, Probably Asian countries, for example, are really underrepresented in the 23andMe database. So, therefore, I think this will have an effect on your ancestry results, but I'm not actually sure about whether it would mix.
Turkish with German people
Benjamin James Kuper-Smith: I just heard, I mean this was a while ago, so maybe they've corrected for this or whatever it's, yeah, it doesn't matter. Anything interesting in yours? Did you, I guess you [00:36:00] did it either presumably for the hope of something interesting and not just 100 percent Dutch or
Laura Wesseldijk: yeah, well, so I thought that I could possibly have some Indonesian roots, and it turned out I did not. So that was a surprise, a little bit disappointing surprise to be honest. And also I had quite a bit of Scandinavian roots, which I didn't expect. There's really nothing Southern European in me.
It's all quite Northern. And I didn't know about that, but no, the main disappointment was the lack of Indonesian No,
Benjamin James Kuper-Smith: maybe, I don't know whether this is interesting, but okay
Laura Wesseldijk: my, my grandfather was born in Indonesia and there were there were a lot of Dutch people in Indonesia
Benjamin James Kuper-Smith: Oh yeah, I was about to say there's a bit of a history there, yeah. Yeah.
Laura Wesseldijk: Indeed.
Benjamin James Kuper-Smith: yeah. Sorry, I forgot about that for a second. So I was like, why are you Yeah, okay.
Laura Wesseldijk: Yeah, I just picked a random country that I wanted to have some genetic [00:37:00] influences from.
Benjamin James Kuper-Smith: yeah. Yeah, no. I think I've also never done it because it would just be Northern Europe. Like, yeah. I don't think it's going
Laura Wesseldijk: Wait, Wait, yeah.
Benjamin James Kuper-Smith: interesting.
Anyway. Yeah.
Laura Wesseldijk: give it a try.
Benjamin James Kuper-Smith: Yeah, I don't know. Yeah, I somehow just don't find it particularly, I just, I guess I just don't expect it to give any interesting results. But maybe that's exactly why I should do it, to be surprised. You mentioned earlier, I mean, that you need these very large sample sizes.
I just had like six to 8, 000 as you had maybe in that Beethoven paper, it being, very large in most studies, but for psychology experiments, neuroscience experiments, but, tiny for some of the studies you do.
Laura Wesseldijk: Wait, so maybe, let me correct you, actually, so before, so the 68, 000 that's in a sample of polygenic scores, and that's Large, but what I was actually referring to is that these GWASs need really large samples, and then we're talking about, so the clap to a beat GWAS [00:38:00] was 600, 000 genotyped data from 23andMe, and we need millions of people.
So for GWASs, we're really talking about millions of people. And then this 6 to 8, 000 sample That's actually in post GWAS analysis. And those can be smaller. Those can be around a thousand or, I mean, that's less, it's the GWAS that needs to really large sample.
Benjamin James Kuper-Smith: Okay, I guess, okay, then my question is specific to, to the GBA studies, which is like almost, can you test anything interesting with a study like that, given that you have to have millions of people who have done who you have data on something like, high, I mean, there's probably a reason why they do heights, just because it's easy to report and everyone knows it.
The musical clapping thing, I mean, I guess we can go a little bit later into that task in particular, but if I remember correctly, it was self reported and, it's, it just seems to me like you're bound to very low resolution data that, or, yes, I'm just curious, like, can, [00:39:00] are these studies then, just because you need so many people, just naturally limited in terms of the I mean, how interesting it is a bit of a judgmental question, but it seems to me that you can only ask very basic, very easy to measure things with those studies then.
Or is there any way around that potentially?
Laura Wesseldijk: No, you're completely right. This is a very good question. Of course, when you're talking about samples this large the measurements become quite simplistic. Ideally the measurements would be way better. And this would also improve predictive value of polygenic scores then later however, what what we did show this clapping to a beat G was, which is obviously a very limited measurement of musicality.
Because it has such a large sample, 600, 000 people were genotyped and their genotype was analyzed and we knew whether they could clap to your beat, yes or no. If after constructing polygenic scores based [00:40:00] on this gwas, we did actually see that it taps into musicality. The polygenic scores do predict with small effect sizes, but significantly a broad range of musical aspects.
So we really. showed also with follow up studies that even though the measurement is not optimal, you are touching upon something informative there and it can be used to further investigate gene environmental interplay underlying musicality and the development of musicality, for example. So I think what GWASs have shown now over the last decades that it is very important to, um, yeah, improve sample sizes, that is more important than to focus on a small group with a great measurement.
However, ideally, and this is of course all, Work that is happening right [00:41:00] now both the measurement and the sample size will be improved. Especially when you consider musicality, which is just such a complex trade that consists of so many different aspects. But also in other psychiatric disorders, for example, for when you look at depression, it's so heterogeneous.
You can be depressed and sleep all day or not sleep at all eat a lot, not eat at all. Like if you're going to run a GWAS on having depression, yes or no, you're tapping into so many different characteristics there. So it is very important, but I think overall, what is most important is to have large sample sizes, and then well, I mean, I can't say what is most important, to be honest.
I mean, it's both very important, but a lot of large GWASs on which you could call quite imprecise phenotypes have shed light on [00:42:00] mechanisms and genetic variants that have been informative But are obviously not optimal.
Benjamin James Kuper-Smith: Okay. I wanted to ask a little bit about the clapping to the beat task or questionnaire or whatever. Is that is that a test of musicality really, or just of like motor control and working memory?
Laura Wesseldijk: The authors call it beat synchronization. It's yeah, actually in the original GWAS, they call it beat synchronization and not musicality more indeed related to motor control. However, a followup study, which was conducted by me and my colleagues showed actually that when you. Make use of this polygenic score, it really predicts musicality outcomes or musicality.
Not so much musicality, just a general interest in music or an enjoyment of music. It predicted. The amount you listen to [00:43:00] music, whether you play an instrument the start age whether you practice music then it did predict, it did predict musicality tests such as rhythm ability, melody discrimination ability, and pitch discrimination ability.
So some objectively measured music tests were predicted by the. Genetic predisposition for beat synchronization. And actually, interestingly, not at all things such as sport behavior or intelligence or other type of creative achievements. So we actually concluded, well, it taps into a general interest and enjoyment of music.
However I would say that of course, whether you can clap to a beat, yes or no, it's a self report. Which isn't ideal, and then it's clapping to a beat it's quite limited.
Benjamin James Kuper-Smith: Yeah, I'm really surprised that beat that this beat synchronization I mean, [00:44:00] not that it correlates necessarily with musical stuff, but that it doesn't correlate with sporting stuff because it seems to me that , yeah, of course you do need very precise, uh, how do I say it,
Laura Wesseldijk: Oh wait, so this question literally was, can you clap in time with a musical beat? So,
Benjamin James Kuper-Smith: the least fine, like it's like the very basics of, that's what you'd expect like a two year old to be able to do, to A or B, to be able to clap with it. And so I'm really surprised, yeah, that has anything to do with musical ability and not sports or something.
Laura Wesseldijk: No, but I mean, if it would be, I mean, it is, quite clearly musical beat. So I think what happens also is that people that are interested in music like music or played music would say yes to this and people that are less involved with music would say no and possibly more likely interpret this as a music related question than a beat or motor characteristics [00:45:00] type of
Benjamin James Kuper-Smith: Yeah, so it's not can you clap in time to the clock, but okay. So do you think maybe the question actually is often interpreted? Yeah, it's almost more musical question because it's because of the framing of the question rather than the actual like activity. It's supposed to represent.
Laura Wesseldijk: Indeed. And I'm now, I'm gonna, I'm trying to search for you how many people actually said no.
Benjamin James Kuper-Smith: Wasn't it like 10%? Again,
So
Laura Wesseldijk: 555, 000 people that said yes, and around 50, 000 people that said no to the question can you clap in time with a musical beat?
Benjamin James Kuper-Smith: 9 percent Yeah, I mean, that's, I'm surprised it's that many people. I mean,
Laura Wesseldijk: Yes.
Benjamin James Kuper-Smith: maybe they were thinking of complex jazz rhythms. I don't know. But yeah, anyway, I guess I don't want to judge people can't clap to the beat, or just
Laura Wesseldijk: but I actually, I would be more worried about [00:46:00] people saying yes to this question while they can't
Benjamin James Kuper-Smith: Yeah.
Laura Wesseldijk: clap to a beat. I think there's a bit of a bias there as well. I have tested it on a couple of friends who really cannot clap in time to a beat.
Benjamin James Kuper-Smith: Okay. So but I mean, this is obviously like a topic here is The difficulty of measuring, the thing you want to measure, because in this case you have, you need lots of, I mean, even in like the smaller studies, you can't just do like 30 people, I'm assuming. So you need lots of people, and you need, I mean, for musicality, is it, I'm assuming it's all self reported, or do you actually do studies where they actually have to perform something, and it's measured, and
Laura Wesseldijk: Yeah, so where I started my postdoc at the Karolinska Institute in Stockholm, and there's the Swedish Twin Registry, and there they have a cohort where they heavily phenotyped the twins on, Music related traits and among all those questions [00:47:00] was also a test an online test that the twins completed.
So around eleven and a half thousand twins completed the rhythm, melody, and pitch discrimination test. And this is actually I think, I hope I'm correct, but I think to my knowledge, this is the only large twin sample in the world that has actually got twin data on measured or objectively measured data.
Musicality. So where you actually can conduct genetic analysis based on because yeah, the majority of twin data or genotype data are based on self-reported measures of music related behavior. Very often, not even musical achievement, level of musical achievement, but more how many hours did you practice?
Did you ever play an instrument? Those type of questions.
Benjamin James Kuper-Smith: So just briefly, 11, 000, 11, 000 Swedish twins, that has to be like most of [00:48:00] them, right? Because there's only like 5 million Swedish people and twins aren't super That must be like a Do they just do this with all the twins there? Because it can't be like that many more.
Laura Wesseldijk: there are many more. There are many more. Yes. Yes. I don't know. I don't know how many Swedish twin register. 87,
Benjamin James Kuper-Smith: million Swedish people and there's like, what, like 1 percent is twins or something I don't know.
Laura Wesseldijk: twin pairs
Benjamin James Kuper-Smith: See, there you go, but lots of them are very old. Oh, in the, 78, wow.
Laura Wesseldijk: No, sorry.
Benjamin James Kuper-Smith: Wait, the,
Laura Wesseldijk: I'm saying it wrong. I don't know. 87,000 twin pairs in the Swedish twin registry. 87, 8 and seven.
Benjamin James Kuper-Smith: But that has to be, but not all of those have the, did all the testing or
Laura Wesseldijk: No. So this was only twins were approached from a certain cohort, an adult cohort. It's called the stage cohort of [00:49:00] the Swedish Twin Registry. So they were around, yeah, between 30 and 50 years old. Average 40 years. They're different. Cohorts in these twin registries. There are many all over the world.
I did my PhD in Amsterdam at the Netherlands twin registry. Which I believe is even larger than than the Swedish. So they, they have different cohorts. They have children's cohorts. They have twins that they follow up from birth. They have adult cohorts. They have aging cohorts. They're a lot of different Databases cohorts that you can.
Yes,
Benjamin James Kuper-Smith: one, there were so many twins, and number two, that they so systematically get, get tracked and measured by these geneticists.
Laura Wesseldijk: They're very important. We uh,
Benjamin James Kuper-Smith: time job for them.
Laura Wesseldijk: we need twins. So no, it's quite actively twins are approached. In the Netherlands at birth, when you give birth to twins, you immediately will receive an invitation to participate in the twin register. [00:50:00] And also there's active recruitment among adult twins. So for example, these are organized where you can go as a twin.
And then you're shown results from scientific work that is based on twins. And of course ask, do you want to participate in certain questionnaires or tests?
Benjamin James Kuper-Smith: Okay. Yeah. Now I'm trying to think of a triplet study, but I guess it
Laura Wesseldijk: They're also there. The
Benjamin James Kuper-Smith: Does that make
Laura Wesseldijk: also has a triplet cohort.
Benjamin James Kuper-Smith: Okay, does that make Does that give you anything that the twin doesn't already do?
Laura Wesseldijk: No, it just gives you extra data.
Benjamin James Kuper-Smith: Just sounds fun. Okay. Let's see. Yeah, I mean, so you have an article on well, there's a big question, I guess, in music, whether there's a critical period to learning music, that kind of stuff, you have an article on it, um, yeah, maybe do you just want to, again, outline kind of, yeah, briefly as a kind of first overview of what you did and then we can jump into individual points.
Laura Wesseldijk: Yes. Yeah. [00:51:00] So, I think this paper is actually a great example of how you can utilize twin data beyond estimating heritability which twin studies are most well known for, but it's possibly not super informative to just know, okay, genetic factors are of play, but in this paper, so there's the idea That you should start practicing music at a young age, if you want to become a really successful musician.
And it is true that when we look at successful musicians, there is an association between the age they started and the younger people start the higher the risks that they are either professional musicians or successful musicians reach higher levels.
Benjamin James Kuper-Smith: Risk is a nice word though. But
Laura Wesseldijk: Did I say risk?
Benjamin James Kuper-Smith: I think you said they're the higher risk that they become professional musicians. No, but actually, I mean, just just under, underline that point. I mean, especially, I think it depends obviously what type of music you do. I think in pop music, I think it doesn't matter that much, but in classical [00:52:00] music, which is what I know most about, I mean, for example, with classical pianists, I only know of one pianist who started like after the age of eight or something like that.
Interestingly, his surname is Wunder, which is miracle in German. He's Austrian. But basically, and I've heard a lot of professional musicians say, I think Daniel Barenboim once said, like, if you start after a certain age, you just, I mean, interestingly he, he said you can't become world class, I think because of the, the, the dexterity in the fingers, it wasn't musicality per se, it was more like you're just not going to get the smoothness in the movements you can make and that kind of stuff if you don't start very early.
But I mean, certainly in classical music, there's basically, as far as I'm aware, no examples of people who started, I mean, after eight, basically. So it's from my world it's basically dogma in that sense. Yeah.
Laura Wesseldijk: Yes, indeed. And the idea is that, yeah, there is some kind of sensitive period that you shouldn't miss out on for the development of these skills that you [00:53:00] would need.
So it's true. There is an association between the age you start practicing music and the level of music that you will reach in adulthood. And but obviously there is an effective practice. Because the younger you start, the more hours of practice you will accumulate. So what we did is we looked at a professional musician sample of around 300 professional musicians and a twin sample of around, I think 10, 000 twins.
And we saw indeed that in both cases, in both samples, professional musicians and just individuals in a population, the younger you start, the higher your levels of musical achievement in adulthood. But then when we control for practice hours, this association only remains when we look at objectively measured music skills in adulthood, but not anymore, the levels of [00:54:00] music that you reach, whether you become a professional musician or an amateur musician, for example, so that association is really driven by practice. And then if you want to really know whether the age you start, whether you have to start before a certain age, because otherwise you miss out on a sensitive period, whether it's really causal, the age you start on a later musicality in adulthood, you would expect that the identical twin that started at a younger age, would have higher skills in adulthood than his or her co twin that started at a later age.
Because identical twins, they share 100 percent of their DNA, they share their family, rearing environment. If you expect causality, then it cannot be due to these familial factors such as genes and the family environment. It should really be the age you start. And we saw [00:55:00] actually that If you compare monozygotic twins, identical twins that differ, one started a lot earlier than the other, there is no difference in their musical skills in adulthood.
So it's not necessarily the age you start. There is a relationship. Yes, it is true. The younger you start, the better your skills and the higher your musical achievement in adulthood. However, this is largely driven by genetic and family environmental factors. And it's not so strange if you think about it, because people with a genetic predisposition or talent or a family environment in which they are provided with a lot of music related traits and behavior, they just often start at an earlier age because Bach, for example, it said he started at the age of three or something, but he grew up in a family with only musicians.
He probably had, obviously had. a genetic predisposition [00:56:00] for musicality. Of course, these people search for music, for instruments. So yes, the best, the greatest musicians started at a young age, but whether it's really this particular age causing them to be the best, we can't say that this is causal because otherwise we would have to see it within identical twins that differ in the age they started, some that started before the age of eight while their co twin started after.
And that didn't, turned out to not make a difference. So there really is quite a bit of familial confounding there. And that's just important to keep in mind when you're looking at phenomenal like this, when you try to advise people start before the age of eight. And I'm not saying indeed that we should all start late.
However, it seems to imply that there isn't a cutoff that you need to reach and afterwards don't give it a try.
Benjamin James Kuper-Smith: [00:57:00] I mean, it seems to me that also one differentiation that maybe needs to be made here is that between, being able to do it and being able to be one of the best in the world. Because, I mean, most people are not going to be professional positions. So obviously like you can, if you have fun with it, then you can, still do it.
I guess like, especially when it, when they say like you, you should start before the age of X. I mean that probably also means like if you want to be so good that you can make a living in a hyper competitive field from this I'm just curious like how do you differentiate like think about these two different things because it seems to me that we're often interested in the extreme outliers, but the studies you can do by definition are between people who are one or two standard deviations, maybe three, outside of the norm.
But what we're often interested in is, the freaks that says the people who are just like extremely like you, who make the other professionals look like amateurs. I'm curious, because, yeah, how do you think about this difference in [00:58:00] terms of how extreme the behavior is supposed to be?
Laura Wesseldijk: Yes. So when you make use of twin data, you make use of population based samples. Twins represent the populations represent the population of the country that they live in. So obviously we're looking at on average, yeah, most, a quite average group of people. And therefore in this study, particularly, we also looked within a professional musician sample.
But how many professional musicians will you have in a twin sample? It will be a very small percentage. But this is actually,
Benjamin James Kuper-Smith: briefly, but I guess my point is also that even professional musicians, the average professional musician isn't that great at music, right? Like, let's just be completely honest, right? I don't know exactly what professional meant in this context, I can't remember, but
Laura Wesseldijk: These are people that make a living out of out of making music and performing music. So they get money [00:59:00] for
Benjamin James Kuper-Smith: Yeah, but I guess the point is still, like, it's still, um, I guess it's, I guess the point is more that the extreme outliers by definition are very rare. So you, they're basically, you're not going to get any sample size that you can include in this kind of thing, or they're going to be drowned out by all the other stuff.
And I'm just, I guess the question is what do you, how do you deal with the situation where the absolute extremes that are by definition, but often by far the rarest, that those get drowned out in noise almost by. the bulk of people, even if that's in a professional setting, let's say.
Laura Wesseldijk: Yes, I guess this would be a problem. Yes, how do I do it? To be honest, I don't really know how to answer this
Benjamin James Kuper-Smith: I mean, yeah, it seems to me like this is just a generic problem that any field that I mean, in some sense, maybe it's also not what you're interested in. Right.
Laura Wesseldijk: No, exactly. I was just thinking this will be if I want to know how it is possible that someone [01:00:00] becomes the best and how to reach that. But in the end, we are interested in individual differences and why they occur and why they occur among People and general people not the specific cases, the outliers the special ones.
When we study the development of music acquisition it's with the idea of looking in a general population, who becomes a good musician versus who, who doesn't, but it isn't like, how come this one person completely went so it's more of a, an approach of explaining individual differences than it is zooming into how to become the best ever.
And above all, what you said just previously before, I think also I personally am not so interested in how to become the best musician and what you need, what to do, perfect [01:01:00] genetic and environmental circumstances to become an amazing musician, but also just music engagement in general and enjoying making music.
And a lot of my papers have also been about you don't, I mean, you don't need to become the best. Well, that's not what most of my paper's been about, but more like, okay. Music and mental health, for example, how are they associated? Can it be coping mechanisms to yeah, deal with certain mental health problems, so not necessarily only focusing on expertise development.
Benjamin James Kuper-Smith: Yeah, I mean, I didn't exactly mean it that way, but I meant to almost more in terms of like those people should be. genetically the most, like they should have, I mean, maybe this is just a fallacy that the assumption is that these people should have like all the genetic traits, so like they should, but maybe as in the case of Beethoven, maybe that's not necessarily the case.
I don't know.
Laura Wesseldijk: Well, I understand your question a little bit better now. So I think we know nowadays that there is very [01:02:00] complex gene environmental interplay underlying things. So you can have the genetic predisposition and then if together with the perfect environment certain interactions occur and correlations because you're also provided with this family environment that's musically enriched and then you're provided with music schools and the exact everything is going completely right, then these people will flourish.
But there can be thousands of different ways of interplay. There can also be a different genetic predisposition with a different with a, an amazing environment or a not so amazing environment with a really strong genetic predisposition and interactions may occur at different time points. And so, I mean, it's going to be almost impossible to say what is the perfect set of factors because there'll be different combinations in different amounts [01:03:00] of genetic and environmental factors that will together reach that level.
Benjamin James Kuper-Smith: And I guess in some sense, it's probably also just a huge fallacy in some sense. I mean, it just occurred to me, the example in football, Lionel Messi, he's too short basically, right? He's like, if you were to build the perfect football, it wouldn't, you wouldn't give him that height and all that kind of stuff.
But somehow he, became basically the best of all time. So I guess maybe the idea that there is like one stereotype that's best for something is a bit silly anyway. Okay. Usually I ask a bit about how people got into the different things.
Maybe in your case maybe I'm curious, like how you went from your PhD to your postdoc, because from what I can tell your initial research was not at all on music and then it got into that in part over time because I guess you still do a lot of stuff on mental health.
So yeah, I'm just curious how and why the step to music maybe starting in your PhD and then how that changed.
Laura Wesseldijk: Yeah. Yeah. So I always had a Interest in music from child [01:04:00] years onwards. And I played very actively music. I played the trombone. Then I switched to electric guitar during my adolescence because I really liked rock music. And obviously the trombone wasn't a
Benjamin James Kuper-Smith: cool as the electric guitar.
Laura Wesseldijk: no, exactly. I did quit playing music when I went to university but still.
actively engaged in going to concerts and listening to a broad variety of music. So I really like music. I was, I studied psychology then biological psychology, and I did a PhD in psychiatry and behavioral genetics. So by the end of my PhD, I was yeah, a trained behavioral geneticist. So very good with genetic research methods.
And actually after my PhD, I was a little bit in doubt whether I wanted to pursue a career in academia, but then I got offered a postdoc position at the Karolinska with the Swedish twin registry specifically on [01:05:00] music and on the acquisition of music the development of music acquisition, I mean, sorry.
And Gene environmental interplay. And I thought it was so interesting because music is my favorite topic. So for me, really was a way to yeah, apply the knowledge I have of genetic research methods to the topic I like the most. And the idea was that I would just do this for a year. But it was such a success and I worked together with Frederik Ullen and Mirjam Mosing Who were both already working on the genetics of music and we worked together really well So I just stayed and they extended my contracts until I just became a full part of the team
specializing actually also in music engagement and mental health. So still combining some of my PhD work with music.
Benjamin James Kuper-Smith: But it sounded like I mean, it sounded slightly random the way you [01:06:00] said you were offered the job of doing music. Like, how did that happen? Was it where you did you know them before? And? Or were you just looking for jobs? And that looks interesting? Or?
Laura Wesseldijk: No, I was by a colleague a colleague of mine who is the head of the department in Amsterdam where I physically work the department of psychiatry, he worked previously for them and did twin analysis and she got offered a professor position and she asked me if I would want to take the postdoc. So, yeah. Yeah, it was it was no active applying to this job or they didn't actively search for me. It was all very coincidental. And I just thought music, I love it. Let's do this.
Benjamin James Kuper-Smith: But you hadn't thought about doing that before?
Laura Wesseldijk: Actually, it never really crossed my mind that this was because I was very much in the field of psychiatric genetics. That this was, it was also an option to apply it to different phenotypes.
Benjamin James Kuper-Smith: Okay. But you're still doing Mental health stuff, [01:07:00] right? Like the music is part of what you do now. Is there just not much of a direct connection just because you randomly got off of this job And now you have these two slightly separate strands I mean, I know you do some combination or is it for you really like there is one kind of integrated View of why you do music and mental health and all that kind of
Laura Wesseldijk: so it's not, well, it isn't that random. So also the department of the Karolinska and now it is the Max Planck where I work. Most of the behavioral geneticists have worked in the field of mental health. I work at the department of psychiatry as well and Amsterdam UMC. So there is a lot of research going on the relationship between making music and wellbeing and mental health.
So it wasn't a very new topic. It was just, it's never been analyzed with genetically informative designs, which I hope that by the end of this podcast, now people realize is important and to take into account, because there is a lot [01:08:00] of. genetic variation between people. And if you don't take that into account, you get biased research findings.
So it's for me, the perfect opportunity here to investigate this with making use of twin and genotype data. And we are still, so the the department for which I work. Doing all the music and genetic research is closely linked to the department of psychiatry.
Benjamin James Kuper-Smith: A brief kind of random question that might take out because it might be too boring, but I'm just curious You seem to work at like three different institutions or something like that. And I was very, when I wanted to contact you, I was vaguely confused, which email address to like, which one to find and that kind of stuff.
And I guess I just used the one from the paper. But like practically, I mean, the, from what I understand, Fredrik Olleen and group are now in Frankfurt or something like that. But you're in Amsterdam. And so, yeah, just how does that work practically? Or is it more on
Laura Wesseldijk: therefore I think my previous [01:09:00] answer was a bit weird because it's a little bit like it's all intertwined
Benjamin James Kuper-Smith: Yeah.
Laura Wesseldijk: so I work, my main job is at the Max Planck Institute for Empirical Aesthetics in Frankfurt at the Behavioral Genetics Unit of Miriam Mosing and Ullen is the director of that Max Planck.
But they came from the Karolinska where I did my first postdoc. On the development of music acquisition using genetically informative designs. So they actually, we all together went to the Max Planck in Frankfurt. However I live and am based in Amsterdam. And my desk is at the department of psychiatry because it's actually a psychiatric genetics department.
And this is really where a lot of researchers work that work with the state of the art. genetic methods. So a lot of people that develop methods in the field of psychiatric genetics and that are very You know, [01:10:00] how to work with polygenic scores, all the big data analysis and cleaning and methods.
So it's really convenient for me to be there as I am a behavioral geneticist after all. So it's really a way of it's a great collaboration between the two places so that I can apply everything that is being developed here and everything that I hear about to the topic, which is studied. With them at the Max Planck
Benjamin James Kuper-Smith: Okay.
Laura Wesseldijk: because over there, we're still, well, we're now growing, but we're still a rather small group of people that do behavior genetics in the field of music.
Benjamin James Kuper-Smith: Yeah, I mean, so what is yeah. Simple question, but what are some of the future things you want to do with music? I mean, and genetics. Um,
It at that open. Kind
Laura Wesseldijk: the relationship between listening to music and [01:11:00] mental health and different genres and the amount you listen to music. I have previously quite extensively investigated the genetic relationship between playing an instrument and mental health just general amateur musicians, but also professional musicians.
And I want to stretch it a bit further to more passive music engagement, namely listening to music, because that's what. More people do actually, but also a lot of studies on, for example, gene environment interaction whether certain genetic predispositions for cognitive abilities interact with practice in what levels of musical achievement you will reach or other types of gene environmental interplay.
So what we really do at the behavioral genetics unit of the Max Planck is trying to investigate how exactly. the genes and the environment work together in the development of musicality. So trying to disentangle that [01:12:00] with behavioral genetics methods.
Benjamin James Kuper-Smith: Okay. At the end of each interview, I ask the same three questions.
The first one is, what's a book or a paper that you think more people should read? It can be old, it can be new, it can be famous, it can be completely unknown. Just something that, oh, you think, yeah. This is really cool, people should read it.
Laura Wesseldijk: Yes. I have two books. I've thought about this. They're very related to the topic of this podcast. The first one is by Paige Harden. It's called The Genetic Lottery. And I think it is Yeah, it's a very interesting book, actually, about how yeah, and why DNA matters for social equality.
Because The field of genetics, of course, has been misused a lot in the past, and has a quite horrible history also with eugenics, and I think people get quite scared often when they hear about genetics, and that there's this determinism about it, which, obviously, is incorrect something can be [01:13:00] highly heritable and that doesn't Need to mean anything bad.
I mean, eyesight is highly heritable and I have definitely very bad genes for eyesight, yet I see fine because I wear contacts. So the environment can completely counteract things. So, there's, but there is this fear for genetic research and I think in her book, she really discusses knowledge of genetic science.
And what we know by now and how actually we should make use of it to create more equal societies because equality doesn't necessarily mean that everybody gets the same because we're not all born the same. And if we want to reach equality, we should take into account that there are yeah, genetic differences.
And I think she really writes Very well, and makes a great point. So that's the first book I would recommend for people that are either interested or scared [01:14:00] of genetics. And then there are these books from Adam Rutherford. How to Argue with a Racist is a book that I thought would be interesting to mention.
Because also he's a geneticist as well. And he writes really nice about also The genetic studies and their results and how they do are not in favor of race as yeah. He says race doesn't exist in a biological meaningful sense. And of course there is race as a social construct and that's real.
But in a biological meaningful way it's not there. And I think it's been very eye opening for me to read about this especially in the field of genetics. he also has a great book called Control, about the history of genetics and eugenics. Which was also very interesting if you want to read a bit more about the history.
Benjamin James Kuper-Smith: It's funny, I I guess we've been talking for like an hour about [01:15:00] genes and we didn't even, we didn't even go to the, I didn't even think about that. Yeah. But yeah, I, it's true. That is a big point in these or can be in some of these discussions. Yeah.
Laura Wesseldijk: Yeah, it is and it is very important, I think, especially geneticists and the researchers are also should be very aware of this. And we all have the responsibility also really be careful with how we bring our research findings out to the world to really try and have as least misinterpretation as possible.
Benjamin James Kuper-Smith: Yeah. Just on the other to, to by the series with the silly the I don't like your example with eyesight earlier because everyone in my family has glasses and I don't. I don't
Laura Wesseldijk: Ah,
Benjamin James Kuper-Smith: I perfect vision so far, but it's, it feels like I'm just waiting for it to stop. I have, I think genuinely everyone, all of my ancestors and siblings have glasses and I'm just
Laura Wesseldijk: Oh,
Benjamin James Kuper-Smith: so far so good.
Yeah. Well, again, my point is we'll see how long it lasts. If you say it's highly [01:16:00] genetic.
Laura Wesseldijk: No, because I mean, obviously with, there's random, I mean, this is genetics, not, it can
Benjamin James Kuper-Smith: Well, so yeah, as you said literally five seconds before, it's not deterministic.
Laura Wesseldijk: No, you be happy without glasses and contacts and don't be afraid.
Benjamin James Kuper-Smith: Oh, thank you. Second recurring question is what's something you wish you learned sooner? Could be from your private life, from your work life yeah, just something you learned and maybe also how maybe if you want to, how you learned it or what you did about it.
Laura Wesseldijk: So I think what I would have liked to know earlier during my career and PhD as well, Is that almost everybody has an imposter syndrome. Yeah. Of course, I also had a huge imposter syndrome. Felt stupid half of the time. So much new information coming your way. Methods and I think it's really, it was really comforting for me to know that other people in science academia, probably in [01:17:00] lots of different fields and everywhere in the world struggle with this.
But I thought, especially in academia gave me a kind of reassuring comforting that we all, because the more you learn, the more you study, the more you research, the more you realize how little, you know, uh, and instead of that making me insecure it can also be empowering and just being okay with, okay, there's this whole.
A large pile of things I do not know, but I can learn them and everybody is facing this.
Benjamin James Kuper-Smith: So did that realization in itself help or, yeah,
Laura Wesseldijk: helped. Yes. Yes.
Benjamin James Kuper-Smith: for someone listening. Um, Final question. Is basically wise for people like me who are on the border between being a PhD student, being a postdoc.
You can take this more metaphorically if you want to, but yeah, any advice
Laura Wesseldijk: So you [01:18:00] already have a postdoc, right?
Benjamin James Kuper-Smith: be starting on Tuesday, you have to move
Laura Wesseldijk: going to have it? Sorry. Where are you going to do a postdoc?
Benjamin James Kuper-Smith: uh, Zurich, I'll be doing so I do like, well, I did my PhD was mainly, it was pretty much only behavioral, but in a neuroscience lab too, and obviously could be doing yeah, rough description is decision making in humans with fMRI and TMS.
Laura Wesseldijk: Very nice. Congratulations.
Benjamin James Kuper-Smith: Thank you.
Laura Wesseldijk: So yeah, I wasn't sure whether to interpret this question as what are you going to do in the meantime?
Benjamin James Kuper-Smith: So no, just generic for, I guess for I should say. I mean I guess it's just, especially for lots of people who are, in the middle nearing towards the end of their PhD, is there anything you'd give advice for them for how to think about what to do or I don't know.
Laura Wesseldijk: my advice would be, which helped me personally a lot to really, Do research studies and projects that you like, that you intrinsically are [01:19:00] motivated about. For me, that has been a huge switch when I switched from psychiatric genetics during the PhD to genetics of musicality, because of my interest in music.
And I realized, Over the last year is that by doing something that I really like and I'm so interested in and makes me happy to go to work every day and makes me, keeps me curious, just by doing the things I like, I am a better researcher, I get more jobs, I get more, I do better studies I am approached more for with collaborations.
It has really helped me to grow as a researcher and to become a successful researcher and to grow in academia by actually doing what I like. And this sounds Almost a little bit corny and ridiculous, but I mean, often in science you're pushed into a project or work that could either be good for your [01:20:00] career or is important for somebody else.
And and of course, I'm not saying you should be very selfish and only do what you want, but it, for me, it has been really beneficial to choose you. Topics and work that I like and that I am curious about.
Benjamin James Kuper-Smith: Yeah. And I guess, I mean, to add here that, a few people who have asked this question to you on the podcast have said, postdoc is a great time to make a switch if you want to make one. So if I guess that's exactly what you did, maybe not as intentionally uh, but or planned at least, but yeah, I guess, if there is a discrepancy between what people like doing and what they are doing, this would be a good time to make a switch.
Some sort of change to that.
Laura Wesseldijk: Yes. Yeah.
Benjamin James Kuper-Smith: great. Well with that, thank you very much and,
Laura Wesseldijk: You're welcome