Taking a lifespan approach to polygenic scores

Abstract

This commentary is a call to action for researchers to create and use genome-wide association studies (GWASs) with previously missed age groups (e.g., infancy, elderly), which will improve our ability to ask important developmental questions using genetic data to trace pathways across the lifespan.

In the target article, Burt challenges the “sociogenomics revolution,” which is thriving because of the incorporation of polygenic scores (PGSs) into social science research. The ease of using PGSs is tempting; however, a source of risk exists in the confounding of genetic and environmental influences because of a variety of biological and statistical reasons. Here, we argue that there is another important confound to PGS work that was overlooked by the author: The lack of consideration of genome-wide association studies (GWASs) in the context of development. Currently, most GWASs with large sample sizes have focused on identifying risk alleles associated with outcomes in adulthood. A consequence of the lack of developmental work in this area, together with the fact that a person's genes do not change over the course of their lifetime, is the assumption that adult GWASs can and must be used to infer outcomes across the lifespan (Harden & Koellinger, 2020). This assumption becomes especially problematic when studying developmental traits, because the manifestation of underlying characteristics changes over the lifespan (Martin, Ressler, Binder, & Nemeroff, 2009).

One example of a disorder that can change in symptoms and forms across the lifespan is anxiety. Specific phobias often predominate in childhood, social anxiety increases in adolescence, panic disorder becomes more common in adulthood, and worry disorders often occur in older adults (Lenze & Wetherell, 2022). Although anxiety disorders are often comorbid and there are transdiagnostic traits shared across these anxiety disorders, there are also characteristics that are unique to each different disorder. The dynamic sets of symptoms associated with psychopathologies such as anxiety can lead to a variety of outcomes after an individual receives a diagnosis, with the associated behaviors becoming more extensive and chronic or the attenuation of symptoms leading to no longer meeting criteria for the disorder (Bystritsky, Khalsa, Cameron, & Schiffman, 2013). These diverging trajectories of psychological disorders may be because of a variety of genetic and environmental factors. As a result, evidence from multiple longitudinal studies in this area supports a “developmental dynamic pattern” in which there is heterogeneity in developmental trajectories of symptoms and phenotypes across the life span (Martin et al., 2009). Using this model, as opposed to the “developmental stable model” in which genetics is thought to be mediated by one unchanging set of risk factors (Martin et al., 2009), is essential to accurately contextualize PGS studies.

The notion of dynamic genetic patterns is changing the way we approach studies of developmental traits, and this approach has been highlighted in studies on attention-deficit/hyperactivity disorder (ADHD) (Rovira et al., 2020), body mass index (BMI) (Couto Alves et al., 2019), and asthma (Pividori, Schoettler, Nicolae, Ober, & Im, 2019). Each of these studies calculated PGSs to investigate the genetic architecture underlying the trajectory of certain risk factors using data from infancy and childhood and found that the genes underlying these outcomes differed over time. More specifically, ADHD possesses a different set of genes that predict the onset and persistence of the disorder (Faraone & Larsson, 2019); BMI possesses heterogeneity at the LEPR/LEPROT gene, revealing longitudinal variation in BMI for infants versus children (Couto Alves et al., 2019); and asthma shows age-related changes across multiple points in the genome (23 genes were childhood-onset specific, one was adult-onset specific, and 37 were related to both childhood- and adult-onset asthma) (Pividori et al., 2019). These findings highlight the importance of the inclusion of wider age populations in this line of work to gain a holistic understanding of the biology underlying developmental outcomes (Couto Alves et al., 2019; Pividori et al., 2019; Rovira et al., 2020). By including age as a covariate, we can map the pathways by which genetic risk manifests across development, and thus study more effectively how various environments and interventions moderate early behavioral manifestations of risk across developmental stages (Dick et al., 2018).

Moreover, the lack of developmental work and provision of age metadata within GWASs (similar to what is being done for sex assigned at birth; Liu, Schaub, Sirota, & Butte, 2012) represents a missed opportunity to ask important questions related to stability and change in biological underpinnings of disorders over time. Other related metadata features can be similarly explored to answer major questions in the developmental field. Using developmentally informed PGSs allows us to ask important questions related to differential susceptibility, such as exploring how the interaction between low socioeconomic status and polygenic risk predicts mental health outcomes or enhancing our understanding of the effects of prenatal supplements on children's mental development (Colombo et al., 2004; Morgan, Shaw, & Olino, 2012). Deepening our understanding of previously well-established biological connections and other developmentally dynamic processes, such as epigenetics, holds promise as an illuminating direction for the field (Shulman & Elkon, 2021). In practice, this can help inform the development of interventions and treatments for individuals with genetic disorders or genetic risk factors (Dick et al., 2018). This work is a call to action for researchers to create and use GWASs with previously missed populations (e.g., early and late in life), which will improve our ability to ask important developmental questions and have a better understanding of how and why certain phenotypes change across the lifespan.