. 2021 Apr 1;11:197. doi: 10.1038/s41398-021-01300-2

Table 1.

Overview of current designs that can be used to study mechanisms of transmission underlying associations between parental characteristics and offspring outcomes.

Design, reference	Genetic transmission	Environmental transmission	Gene–environment correlation	Advantages	Disadvantages
Behavioural genetics designs
Adoption⁵	Association between a biological parent and their adopted-away offspring (shared genes only) indicates genetic transmission	Adoptee’s relative method: association between a parent and their adoptive offspring (rearing only) indicates environmental transmission Adoptee study method/siblings-reared-apart: higher correlation between a biological parent and their lived-with offspring (genes plus rearing) than their adopted- away offspring (genes only) indicates environmental transmission	Higher correlation between biological and living-together parents and offspring (genes plus rearing) than adoptive parents and offspring (rearing only) suggests passive rGE Trait correlation between biological parents and their adopted-away offspring (shared genes only) indicates genetic liability, and subsequent adoptee correlation with the environment provided by their adoptive parent suggests evocative rGE	- If adoption occurs at birth, passive rGE influences (on factors outside of the intrauterine environment) can be excluded as biological parents would have no rearing effect on the offspring	- Generalisability to the general population could be limited, as adoptees may have a higher risk of experiencing a suboptimal prenatal environment - Samples can be difficult to obtain and are usually small - Non-random process of adoption may introduce selection bias - Increase in open adoption (contact between biological and adoptive families) confounds the design
Assisted conception⁷	Higher correlation between a genetically related birth mother (e.g. homologous in vitro fertilisation or sperm donation) and her offspring (genes plus prenatal environment) than a genetically unrelated birth mother and her offspring (prenatal environment only) indicates genetic transmission	Association between a genetically unrelated birth mother (e.g. egg, oocyte or embryo donation, surrogacy) and her offspring (prenatal environment only) indicates environmental transmission	Not studied	- Effective for testing short and long-term effects of the prenatal environment	- Samples can be difficult to obtain and are usually small - Generalisability to the general population could be limited - Prenatal behaviours of mothers who use assisted conception may introduce selection bias - Samples are generally very heterogeneous - Inclusion of families with a within-family donation would bias the design - Design is not optimal for investigating gene–environment correlations
Triparental family (offspring-focused: multiple parental relationships of one offspring)⁸	Association between an offspring and their not-lived-with biological parent (genes only) indicates genetic transmission	Association between an offspring and their step-parent (rearing only) indicates environmental transmission	Higher offspring correlation with their lived-with biological parent (genes plus rearing) than with their step-parent (rearing only) suggests passive rGE Offspring correlation with their not-lived-with biological parent (shared genes only) indicates genetic liability, and subsequent offspring correlation with the environment provided by their step-parent suggests evocative rGE	- Representative of the general population as all types of parent–offspring relationships are included - Large sample sizes can be attained	- Contact with not-lived-with parents can upwardly bias estimate of genetic influences due to passive rGE - Databases with details of family structure are rare
Multiple parenting relationships (parent-focused; multiple offspring relationships of one parent)⁹	Association between a parent and their not-lived-with biological offspring (genes only) indicates genetic transmission	Association between a parent and their step-child (rearing only) indicates environmental transmission	Higher parental correlation with their lived-with biological children (genes plus rearing) than with their step-children (rearing only) suggests passive rGE	- Representative of the general population as all types of parent–offspring relationships are included - Large sample sizes can be attained	- Contact with not-lived-with parents can upwardly bias estimate of genetic influences due to passive rGE - Databases with details of family structure are rare - Cannot investigate evocative rGE as for each child in this design, information from only one parent is known
Children-of-twins⁶	Higher monozygotic–avuncular correlation (between MZ twin uncle/aunt and niece/nephew; 50% shared genes) than dizygotic–avuncular correlation (25% shared genes) indicates genetic transmission	Higher parent–child correlation (genes plus rearing) than monozygotic avuncular correlations (genes only) indicates environmental transmission	If a parental characteristic is largely estimated as heritable (under the effect of genes) in a parent-based twin sample but is under the influence of the shared environment in a child-based twin sample, this suggests passive rGE Estimation of a parental characteristic as heritable (under the influence of genest) in a child-based twin sample suggests evocative rGE	- Can determine if the familial correlation is due to genetic or environmental factors - Extended children-of-twins studies can incorporate siblings and other members of the pedigree and estimate additional parameters	- Samples can be difficult to obtain - Assumes that the size of the genetic contribution to variation in parent and offspring phenotype is the same - Assumes that the same genes influence the phenotype in both the parent and offspring generation
Extended twin (twins and their parents)¹²²	Not studied, as genetic transmission is not estimated but fixed to 0.5 (50% of genes are passed on from parent to child) in the model*	The correlation between parental and offspring phenotype indicates cultural (i.e. environmental) transmission - this captures part of the shared environment effect that is explained by parent-to-child transmission	Covariance between the additive genetic effect and parental transmission suggests passive rGE	- Powerful for estimating shared environmental effects of a specific parental trait that arise due to cultural transmission or social homogamy - Design can be used to study the impact of other family relationships, including siblings - Design can be used to estimate twin-based heritability	- Cultural transmission can be easily underestimated if assumptions of the design are violated or the study is underpowered
Matched-pairs designs
Sibling comparison¹⁰	Not studied, as the familial resemblance between full siblings could be due to genetic or environmental factors	Comparison of outcomes in children with a specific parental exposure and their unexposed full sibling who is otherwise naturally matched for familial (genetic and environmental) risk; higher outcome levels in exposed than unexposed siblings indicates environmental transmission	Not studied	- Generally excludes passive rGE as siblings typically share the same parentally provided environment - Can exclude evocative rGE within the design if certain that the parental exposure precedes offspring outcome	- Requires differential exposure between siblings, which can elicit selection bias - Cannot distinguish if the familial resemblance between siblings is due to genetic or environmental factors - Design is not optimal for investigating gene–environment correlations
Case–control¹¹	Not studied, as cases and control parent–offspring pairs are matched on genetic risk	Parent-offspring pairs are manually matched on familial and genetic risk. Outcomes are compared between children with a specific parental exposure (cases) and unexposed children (controls); higher outcome levels in cases than controls indicates environmental transmission	Not studied	- Representative of the general population - If matched well, ensures no effect of confounding factors	- Matching is done by the researcher and is susceptible to errors - Resources required to find matched parent–offspring pairs - Cannot investigate genetic transmission or gene–environment correlation
Molecular genetics designs
Within-family PGS:genetic sensitivity analysis¹³	The disappearance of an observed parent–offspring correlation after adjusting for offspring PGS for the predictor and outcome traits indicates genetic transmission	The remaining parent–offspring correlation, after adjusting for offspring PGS for the predictor and outcome traits, estimates environmental transmission	Reduction of parent–offspring correlation after adjusting for offspring PGS suggests passive rGE	- Can test whether parent–offspring associations are partly due to shared genes	- PGS capture only a small proportion of heritability and cannot index the effect of all shared genes - Requires well-powered GWAS summary statistics
Within-family PGS: genetic nurture^14,83	Association between PGS based on transmitted parental genes and offspring outcome indicates genetic transmission	Transmitted/non-transmitted method: association between PGS based on non-transmitted parental genes and offspring outcome indicates genetic nurture Statistical control method: association between parental PGS and offspring outcome, after adjusting for offspring PGS to account for shared parent–child genetic effects indicates genetic nurture	Association between offspring PGS and parenting suggests passive rGE Association of offspring PGS with parenting, after adjusting for parental PGS suggests evocative rGE	- Can examine environmental transmission without parental phenotypic information	- Requires well-powered GWAS summary statistics - Datasets with parent–offspring genotyped duos or trios are rare
Maternal-effects genome-wide complex trait analysis (M-GCTA)¹⁶	Not studied*	The estimated effect of maternal or paternal genetic nurture: variance in offspring outcome that is explained by the effect of maternal or paternal genotype (after accounting for transmitted genetic effects)	Covariance between direct genetic effect and genetic nurturing effect suggests passive rGE	- Can estimate the overall impact of genetic nurture from mother or father - Representative of the general population - Design can be used to estimate SNP-based heritability	- Cannot model both maternal and paternal genetic nurture effects at the same time - Large sample sizes are required to estimate multiple variance components based on genetic data - Datasets with parent–offspring genotyped duos or trios are rare
Relatedness disequilibrium regression⁸⁷	Not studied*	The overall estimated effect of parental genetic nurture: variance in offspring outcome that is explained by the effect of mid-parent genotype (after accounting for transmitted genetic effects)	Covariance between direct genetic effect and genetic nurturing effect suggests passive rGE	- Can estimate the overall impact of genetic nurture from both parents combined - Representative of the general population - Design can be used to estimate SNP-based heritability	- Assumes that maternal and paternal genetic effects are the same and of equal magnitude - Large sample sizes are required to estimate multiple variance components based on genetic data - Datasets with parent–offspring genotyped trios are rare
Trio-GCTA¹⁸	Not studied*	The Eestimated effect of maternal and paternal genetic nurture: variance in offspring outcome that is separately explained by the indirect effect of maternal and paternal genotype (after accounting for transmitted genetic effects)	Covariance between direct genetic effect and genetic nurturing effect suggests passive rGE	- Can estimate the individual impact of genetic nurture from both parents in the same model - Representative of the general population - Design can be used to estimate SNP-based heritability	- Large sample sizes are required to estimate multiple variance components based on genetic data - Datasets with parent–offspring genotyped duos or trios are rare

rGE gene–environment correlation, PGS polygenic scores, SNP-based heritability variance in a target trait that is explained by the additive genetic effect of common genetic variants known as single-nucleotide polymorphisms.

*These designs can be used to estimate twin or SNP-based heritability for offspring outcomes, i.e. the proportion of variance in a phenotype that can be explained by genetic variation in the population under study. This does not directly index genetic transmission, although it is implicitly known that children receive their genes from their parents.