Understanding Images: How genetic makeup of a ‘roommate’ can influence health
Author: Amelie Baud, European Molecular Biology Laboratory, European Bioinformatics Institute, Wellcome Genome Campus, Hinxton, United Kingdom
Competing Interests: Amelie Baud is an author of the article discussed in this blog.
Acknowledgements: I would like to thank Mary Todd Bergman and Oliver Stegle for helpful comments on a draft of this blog.
Image Credit: Illustration by Spencer Phillips, EMBL-EBI.
Social genetic effects
We know that people who spend time together influence each other. However, we have a very tenuous grasp of the magnitude and mechanisms of social effects. A better understanding of the role of the social environment in health and disease is essential, as it may open new avenues for human welfare and medicine.
In the featured article from PLOS Genetics’ January issue, we found that when two individuals interact, the state of health of one can be affected by genes of the other. This brings a new perspective to our work in genetics, and perhaps to the clinical setting as well.
A measured approach
To illustrate how social genetic effects work, let’s consider the way people who live together influence each other.
For example, say that my partner’s genes contribute to making him a very relaxed person who cooks well and has good looks. Given the right circumstances, living with such a person would help me get over bad days. Thus, my partner’s genes would indirectly contribute to my resilience.
This example is simplistic but it highlights the fact that there is a vast number of ways (behaviours, skills, other characteristics) whereby a partner can exert influence on one’s state. This is why it has been so difficult to study social effects so far.
Fortunately, there is an alternative to having to guess which traits of a partner influence the phenotype of interest (e.g. a disease) and measuring them. As those ‘influencing’ traits are likely to be under some level of genetic control, we can measure the genotypes of the partner and use them as a proxy for the partner’s ‘influencing’ traits.
Genotypes are easy to measure in a comprehensive manner, making this an attractive alternative, but it will only work if the indirect associations between genotypes of one individual and traits of another are strong enough to be detected.
To address this question, we looked at more than 200 traits relevant to human health in laboratory mice and quantified the aggregate effect of the genotypes of cage mates. Our results show that anxiety, body weight, immune traits and wound healing were significantly (i.e. beyond chance) affected by social genetic effects. Up to 29% of the variation we observed was attributable to differences in the genetic makeup of cage mates. Furthermore, for several traits the genes of cage mates were more important than the mouse’s own genes.
We saw clearly that the contribution of social genetic effects to phenotypic variation can be large. As social genetic effects capture only the genetic component of social effects, they don’t tell the full story, but they do set us on the right trail. This trail will lead us to stop looking at individuals in isolation and instead study them in their social ecosystem.
Getting to a mechanistic understanding
In our experiments, all kinds of phenotypes were affected –including some behaviours as we might have expected based on common sense. But there were also surprises, such as wound healing. How could this be? Remarkably enough, we were able to detect social effects on healing without prior information on the mechanisms, simply using the genotypes of cage mates.
However, we can reasonably speculate that this influence may be mediated by mice grooming each other, which affects the healing process. Then again, any of myriad stress-inducing social interactions could also have indirectly affected healing (and many other traits), as stress has widespread effects on one’s biology.
Having detected strong aggregate social genetic effects, we are now turning our attention to the specific genes in partners that have social/indirect effects. Establishing these links should enable us to dissect the mechanisms underlying the observed social effects, just as direct genetic effects (effects of an individual’s genes on its own phenotype) have shed light on the “within-body” causes of phenotypic variation.
How it might work in a healthcare setting
Identifying genes of partners that affect the phenotype of interest in a patient has the potential to help human medicine. For example, imagine that you are a morning person and your partner is a night owl. You will probably go to bed every night later than you would like, or your partner might read in bed while you are trying to sleep. Say you develop an illness, and you review potential contributing factors with your doctor. He or she might not ask you about your sleeping habits because they wouldn’t know that it’s relevant to your disease. However, if there was a clear relationship between your disease and specific genes of your partner that are known to control sleeping patterns (e.g. if we had mapped social genetic effects to the Clock gene), then your doctor would know to start looking at the role of your partner’s sleeping habits in your disease. She or he could then advise you about how to make changes that allow you to get the sleep you need to heal.
Since our research was carried out in laboratory mice and not humans, can we expect social genetic effects to contribute to human health and disease? We think so, because they’ve also been detected in other species such as pigs and hens.
However, it is possible that different traits will be affected in humans. Furthermore, we think that different traits will be affected in different individuals, as a result of interactions between genes of the individual and genes of the partner, and indeed we found suggestive evidence for this in one of our experiments.
Will we someday be able to devise a genetic strategy to match people? This is unlikely not only because predicting traits from genotypic data is extremely difficult, but also —and perhaps most importantly —because it is difficult to fully describe what one wants from a partner. Trying to match people based on a few traits is unlikely to outperform existing non-genetic systems for partner selection.
Baud A, Mulligan MK, Casale FP, Ingels JF, Bohl CJ, Callebert J, et al. (2017) Genetic Variation in the Social Environment Contributes to Health and Disease. PLoS Genet 13(1): e1006498. doi:10.1371/journal.pgen.1006498
This article discussed in this post is also studied in the recent PLOS Genetics Perspective article from Benjamin W. Domingue and Daniel W. Belsky, entitled “The social genome: Current findings and implications for the study of human genetics”, which can be read here.