Father absence, age at menarche, and genetic confounding: A replication and extension using a polygenic score

Abstract

Father absence has a small but robust association with earlier age at menarche (AAM), likely reflecting both genetic confounding and an environmental influence on life history strategy development. Studies that have attempted to disambiguate genetic versus environmental contributions to this association have shown conflicting findings, though genomic-based studies have begun to establish the role of gene–environment interplay in the father absence/AAM literature. The purpose of this study was to replicate and extend prior genomic work using the Avon Longitudinal Study of Parents and Children (ALSPAC), a prospective longitudinal cohort study (N = 2,685), by (a) testing if an AAM polygenic score (PGS) could account for the father absence/AAM association, (b) replicating G×E research on lin-28 homolog B (LIN28B) variation and father absence, and (c) testing the G×E hypothesis using the PGS. Results showed that the PGS could not explain the father absence/AAM association and there was no interaction between father absence and the PGS. Findings using LIN28B largely replicated prior work that showed LIN28B variants predicted later AAM in father-present girls, but this AAM-delaying effect was absent or reversed in father-absent girls. Findings are discussed in terms genetic confounding, the unique biological role of LIN28B, and using PGSs for G×E tests.

The influence of harshness and unpredictability on female sexual development: Addressing gene–environment interplay using a polygenic score

“…However, research by Sung et al (2016) who, using a measure of unpredictability that also included residential and employment changes, also did not find an association between unpredictability and AAM. Indeed, a prior study using the ALSPAC data, the focal sample of the current study, did detect the predicted father absence/AAM association (Schlomer & Marceau, 2021). What could account for this difference?…”

“…By adding many variants of small effect together, PGSs can predict more phenotypic variance than any individual variant on its own. Recent primary and replication research using a PGS consisting of AAM predictive variants (Day et al, 2017) found that the father absence/AAM association could not be explained by genetic variability captured in the PGS (Gaydosh et al, 2018;Schlomer & Marceau, 2021); nor was the association moderated by the PGS, reflecting a lack of GxE (Schlomer & Marceau, 2021). Potential genetic confounding or GxE of harshness and unpredictability associations have not been tested using genomic information in general or the AAM PGS specifically.…”

“…Surveys addressed if menstruation had begun and if so at what age. AAM responses from the first available wave were used in the current study to minimize recall bias (Culpin et al, 2014;Gaydosh et al, 2018;Schlomer & Marceau, 2021). Based on the N = 3,011 available cases (82.6%), mean AAM in months was 152.00 (SD = 14.03) and ranged from 91 to 203.…”

The nature of nurture: Darwinian and mendelian perspectives

“…Perhaps it is just that the same genes that affect children's developmental experiences and environmental exposures, including their parenting, also influence pubertal maturation? A number of efforts to at least partially discount this possibility have still found evidence consistent with theoretical expectations upon controlling for maternal age of menarche and GWAS-derived polygenic scores for menarcheal age (D. Gaydosh et al, 2018;Schlomer & Marceau, 2020) and age of first birth (D. . Another strategy that yielded evidence in line with psychosocial acceleration theory used sisters who were full sibs, thus sharing 50% of their genes, finding that, due to a parental breakup, it was younger ones-who spent less time in a fatherpresent family than their older sibs-who sexually matured at a younger age (Tither & Ellis, 2008).…”

Longitudinal Links between Adolescent and Peer Conduct Problems and Moderation by a Sensitivity Genetic Index

“…Nonetheless, this approach is not as statistically powerful, particularly for detecting main effects, as PGS approaches. It should be noted, however, that relatively few studies to date have incorporated PGS approaches into tests of GxE and there are some possible limitations of this approach for GxE (e.g., Schlomer & Marceau, 2021). Most pertinent to the current study is that it may be difficult to detect disordinal interactions (differential susceptibility) relative to ordinal interactions (diathesis stress) using modern PGS approaches given they emphasize main effect detection (see Zhang & Belsky, 2020).…”