Participants
The sample comprised participants from the Avon Longitudinal Study of Parents and Children (ALSPAC), an ongoing population-based study. Pregnant women resident in Avon, UK with expected dates of delivery between 1st April 1991 and 31st December 1992 were invited to take part in the study. 20,248 pregnancies have been identified as being eligible and the initial number of pregnancies enrolled was 14,541. Of the initial pregnancies, there was a total of 14,676 foetuses, resulting in 14,062 live births and 13,988 children who were alive at 1 year of age. When the oldest children were approximately 7 years of age, an attempt was made to bolster the initial sample with eligible cases who had failed to join the study originally. As a result, when considering variables collected from the age of seven onwards (and potentially abstracted from obstetric notes) there are data available for more than the 14,541 pregnancies mentioned above: The number of new pregnancies not in the initial sample (known as Phase I enrolment) that are currently represented in the released data and reflecting enrolment status at the age of 24 is 906, resulting in an additional 913 children being enrolled (456, 262 and 195 recruited during Phases II, III and IV respectively). The phases of enrolment are described in more detail in the cohort profile paper and its update32,33,34. The total sample size for analyses using any data collected after the age of seven is therefore 15,447 pregnancies, resulting in 15,658 foetuses. Of these 14,901 children were alive at 1 year of age. Of the original 14,541 initial pregnancies, 338 were from a woman who had already enrolled with a previous pregnancy, meaning 14,203 unique mothers were initially enrolled in the study. As a result of the additional phases of recruitment, a further 630 women who did not enrol originally have provided data since their child was 7 years of age. This provides a total of 14,833 unique women (G0 mothers) enrolled in ALSPAC as of September 2021. G0 partners were invited to complete questionnaires by the mothers at the start of the study and they were not formally enrolled at that time. 12,113 G0 partners have been in contact with the study by providing data and/or formally enrolling when this started in 2010. 3,807 G0 partners are currently enrolled. The sample for the current study comprises 4,704 mothers and offspring who had complete genetic and parenting data available for the analyses described below.
MeasuresParenting
We used extracted factor scores from derived parenting factors previously reported in Culpin et al. (2020)1which fully details the item section and development of the parenting factors. In brief, potential items were extracted from self-reported questionnaires administered to mothers from pregnancy to age 3 years capturing parenting behaviour, attitudes and knowledge. Items categorised as parental enjoyment, conflictual relationships, and stimulation and teaching (based on parenting taxonomies) were extracted and entered into separate single-factor Confirmatory Factor Analysis (CFA) models. We focused on ages 0–3 years to capture a period of time when most mothers spend time with their children prior to the commencement of nursery school24. The relevant domains for our hypotheses here were enjoyment and conflict. Scores were derived from saved factor scores for each latent factor. A model using CFA to fit the following 3 factors with the following items showed good model fit. The RMSEA (0.024, 95%CI = 0.024–0.025) and the CFI (0.92) indicated that the measurement model fit the data well. More details on the parental enjoyment and conflict factors are available in Supplement 1. To provide predictive validity of the enjoyment factor we looked at how this factor related to the ‘child’s’ retrospective reports of feeling loved as a child, which they reported on at age 24 (n = 1,965). For a one unit increase in the maternal-reported enjoyment score at age 0–3 years, the child is 2.63 times more likely to report that they felt loved at age 24 (95%CI = 1.49 to 4.65, p = 0.001), this remains after adjusting for maternal age, child gender, and other dimensions of parenting such as stimulation.
Inter-partner conflict in the mother-partner relationship
Mothers’ self-report of inter-parental conflict was assessed using the Conflict Tactics Scale Revised (CTS2)35,36. This is a 39-item scale in which partners in a dating, cohabiting or marital relationship engage in psychological and physical attacks on each other and also their use of reasoning or negotiation to deal with conflicts. The questions are designed to be asked about both the participant and the partner which results in two questions for each item, and a total of 78 questions. Higher scores on this scale indicate more reported conflict. The CTS2 has been reported to be a reliable and valid instrument used across different populations and cultures37. In this sample, the mean score on the scale was 21.61 (Standard deviation = 7.25, range = 10–49, n = 10,514).
Polygenic scores for neuroticism
Genotyped data were available on 8,237 children and 8,196 mothers in the ALSPAC study. Data were also available on a small subset of partners in the cohort (n = 1,722). Full details of genotyping and quality control measures are available in Supplement 1, population stratification methods were applied at this stage and analyses are based on individuals of European descent. From the original genome wide association study (GWAS), 116 independent variants were found to be robustly associated with neuroticism7. Of these original variants, 109 were available in ALSPAC. Weighted PGS for neuroticism were calculated for each mother, partner and child in ALSPAC with genetic data using PRSice238. These are calculated as the sum of the number of copies of each effect allele carried by an individual (this ranged from 0 to 2 for each SNP), multiplied by the effect estimate identified in the original GWAS. These weighted sum scores were then standardised prior to use in the analysis. Correlations between PGS from individuals were: mother and child r = 0.49, father and child r = 0.47, mother and father r = 0.05.
Proportion of similarity in neuroticism PGS in relationship pairs
A proportion of similarity measure was calculated by counting the number of neuroticism SNPs at which mothers and offspring, and mothers and partners, had the same genotype, see Table 1. This proportion was then divided by the total number of neuroticism SNPs to give the proportion of neuroticism variants at which mothers and offspring/partner shared a genotype. This proportion of similarity was available for 5,470 mother-offspring pairs and ranged from 0.46 to 0.80. Of these, 4,704 mothers and offspring had complete genetic and parenting data and comprise the sample used for all analyses described below. The partner proportions of similarity were available on a subset of 1,292 couples and ranged from 0.34 to 0.70.
Table 1 Worked example of creation of the similarity score, with 8 SNPS.Data analysis
The sample of mothers with complete genetic data was compared on several demographic variables with those for whom genetic data were missing using chi-squared tests. Associations between PGS, proportion of genetic similarity for neuroticism scores, and maternal reports of enjoyment of parenting, conflict and marital relationship were conducted using linear regression models in Stata version 17. Linear regression was chosen given that all variables were continuous and normally distributed. PGS and similarity scores were continuous exposure variables, and extracted parenting factors and marital relationships were continuous outcome variables. Associations were also examined by child sex.
First, we examined main effects of mother and child PGS for neuroticism on maternal-reported enjoyment and conflict of the mother-child relationship. Next, we examined the association between proportion of similarity in mother-child PGS for neuroticism with maternal-reported enjoyment and conflict in the mother-child relationship. To understand whether the degree of neuroticism (i.e., both members of the dyad being high or both low) might influence maternal reports of enjoyment with their child, we investigated the association between the proportion of similarity between mother and child for neuroticism PGS and maternal enjoyment, stratified by groups according to mother and child both being in the top or bottom 50% for PGS for neuroticism. Next, we examined associations between mother-partner similarity in PGS for neuroticism and maternal-reported enjoyment and conflict in the mother-child relationship. In this model, we also included maternal-reported marital conflict as a confounder in an additional step to examine whether such association might be explained by conflict in the marital relationship.
In addition, we also performed two sensitivity analyses. For our positive control, we calculated a proportion of similarity for neuroticism between mothers and partners and examined the association with maternal-reported inter-partner conflict. Again, to understand whether the degree of neuroticism (i.e., both members of the dyad being high or both low) might influence maternal reports of the inter-partner conflict, we investigated the association between the proportion similarity across all neuroticism SNPs between mother and partner and marital conflict, stratified by groups according to mother and partner both being in the top or bottom 50% for PGS for neuroticism.
As a negative control, we calculated a weighted PGS for Rheumatoid Arthritis (RA) as described above. From the original GWAS, 101 independent variants robustly associated with RA were identified39and of these 87 were available in ALSPAC. We also used these genetic variants to calculate a measure of mother-child proportion of similarity (i.e. the proportion of these SNPs for which mothers and children shared the same genotype). This was available for 5,470 mother-child pairs and ranged from 0.45 to 1.00. We examined the association between the RA similarity measure and maternal reported enjoyment and conflict in the mother-child relationship.