Over the last couple of days, there’s been a great deal of coverage of three new studies on the genetics of schizophrenia. While the coverage has actually been pretty good, almost all the news stories make the same error when talking about the ‘genetic risk’ for the condition.
Twenty years ago, geneticists were searching for the ‘gene for schizophrenia’ until it became apparent that there was not going to be a single gene, or even a handful, found responsible for the mental illness.
It since became a mantra that the genetic risk for schizophrenia would be conferred by ‘many genes of small effect’. In other words, the cumulative effect of lots of genes that, on their own, would be quite benign.
Nature has just published three studies that use the only-recently-feasible technique of scanning the whole genome and has reported the first convincing positive evidence for the ‘many genes of small effect’ theory by finding that a whole bunch of genes, when considered together, account for about a third of the total difference in schizophrenia risk.
Interestingly, all three studies find that many of the genes lie in a <a href="region called the ‘major histocompatibility complex’ – a series of genes involved in the function of the immune system.
However, lots of the news reports, even from science publications give variations on the theme that ‘genetic factors account for 80 percent of the total risk of getting schizophrenia’.
This 80% figure (which can vary, some give 90%) is not an estimate of risk and shows a misunderstanding of estimated heritability taken from twin studies.
Luckily, I tackled exactly this issue in a column for July’s edition of The Psychologist:
Nature versus nurture is a lie. Music is not melody versus rhythm, wine is not grapes versus alcohol and we are not environment versus genes. We are their sum, their product and their expression. They dance together and we are their performance, but neither is an adversary. The art of understanding this elegant ballet is complex and arcane but you may never realise this from reading the quoted results of genetic studies, because the extent to which a trait is heritable, that is, accounted for by genetics, is usually expressed as a simple percentage.
If you search Google for the phrase “80 percent genetic”, you will discover hundreds of sources that claim that everything from schizophrenia, to height, to intelligence has been found to be four fifths ‘genetic’. Pick any other figure and you can find everyone from psychologists, to politicians, to journalists claiming that this or that is explained by genes to a given percentage. Geneticists know the subtly of this percentage and why these statements, usually lifted from the results of twin studies, are misleading, but clearly many others do not.
Imagine a mental illness is described as being 80% heritable. This is often taken to mean that four fifths of an individual’s risk is down to his or her genes, but this is not the case. What it means is that 80% of the variance in the measured illness was explained by genetic factors in the specific group that was studied. If this seems like a frivolous distinction, bear with me, because it is key in understanding heritability and it becomes crystal clear when tackled as an example.
Imagine that we could study a population where everybody lived in an identical environment. They did the same things everyday; they ate identical foods, had identical relationships and were stressed by identical events. Their lives were carbon copies of each other. A twin study would find that mental illness would be close to 100% heritable, because if the environment is fixed, any difference must be down to genetics. In fact, twin studies would find that everything is close to 100% heritable, for exactly the same reason. To flip our thought experiment on its head, if we only studied genetically identical clones, everything would be 0% heritable, because any difference must be down to the environment.
These figures do not necessarily tell us anything about the potential for a trait to be influenced by nature or nurture, because heritability is rarely an immutable and absolute fact about biology; it is an overall measure of how things are for that group, at that moment. In other words, the process of measuring the influence of genetics is, itself, subject to environmental factors. It captures the dance, not the dancers.
Thanks to Jon Sutton, editor of The Psychologist who has kindly agreed for me to publish my column on Mind Hacks as long as I include the following text:
“The Psychologist is sent free to all members of the British Psychological Society (you can join here), or you can subscribe as a non-member by emailing sarsta[at]bps.org.uk”
Link one two three to Nature genetics of schizophrenia studies.
Link to good write up from Science News, despite 80% genetic risk slip-up.
Mind Hacks 
What a lucid explanation and wonderful analogies.
It’s always annoying to me when I read about a new “study” claiming to “prove” something or other, then it turns out the media has completely misread/misinterpreted/misunderstood what the study was actually trying to do. And it’s also very unfortunate that the average person (i e someone with no training in the scientific method) will take what they read in the newspaper as gospel, because they don’t know any better. Your recent post about Steve Connor illustrates that point quite well. Many journalists want to take umbrage when someone points out factual errors in science reporting, and many of them want to take further umbrage when someone suggests that, perhaps, journalists writing about science (or any other specialized field, for that matter) should have a background in science. That’s why it’s such a pleasure to read this post. The media, as you say, tend to want to latch on to that “eighty percent” figure, without stopping to consider what it really *means*. Thank you for making that meaning clear.
I am the daughter of a bipolar father and the mother of a bipolar daughter. I believe mental illness is virtually one hundred percent genetic. If there is an environmental component, from my experience it would be a biological event such as the mother contracting the flu while pregnant. Insisting that there is any percentage whereby nurturing can trump genes leads to blaming the parents. Genes are the blueprint of our bodies. They determine the architecture of our brain, which constructs our personality in all aspects. I have no doubt you can ruin a child and warp personality through extreme abuse. I have seen no compelling evidence nurturing can “cure” or avoid any form of mental illness. My daughter was a magical child, adored and given every advantage. When away at college, at exactly the age most bipolar illness hits, she went crazy. Nine years later she was finally diagnosed as bipolar. It took this long despite my pleading with scores of psychiatrists to consider her family history. Anyone who lives with this has seen the chemical switch flip. My father, a brilliant physician, told me many times that his mental illness had nothing whatsoever to do with his childhood. In every article that argues for the power of environment, I see the only political correctness. The results of twin and adoption studies are conclusive. I refer you to Judith Harris’s excellent book, The Nurture Assumption, as well as all of Steven Pinker.
Beauty is in the eye of the beholder.
The same with mental illness.
Only a Psychiatrist can detect it, no lab test can.
If a lab test CAN then its no longer mental, the mental of mental illness.
Schizophrenia is only an illness/disease/sickness if the person is not a functioning member of society.
To find the DNA for schizophrenia is similar to looking in the DNA for certain type of skin color. As skin color determines criminality of the individual.
“More than three times as many black people live in prison cells as in college dorms, the government said in a report to be released Thursday.”
Schizophrenia is only a disease if authority define it as a disease, just like homosexuality was considered a disease. or maybe blackness.