Everything is heritable: maternal selection of sperm donors for height increases offspring height as much as predicted by heritability estimates.
"Quantitative Genetics in the Postmodern Family of the Donor Sibling Registry", Lee 2013:
"Quantitative genetics is primarily concerned with two subjects: the correlation between relatives and the response to selection. The correlation between relatives is used to determinethe heritability of a trait--the key quantity that addresses the question of nature vs. nurture. Heritability, in turn, is used to predict the response to selection --the maindriver ofimprovementsin crops and livestock. The theory of quantitative genetics has been thoroughly tested and applied in plants and animals, but heritability and selection remain open questions in humans due to limited natural experimental designs. The Donor Sibling Registry (DSR)is an organization that helps individuals conceived as a result of sperm, egg, or embryo donation make contact withgenetically related individuals. Families who conceived children via anonymous sperm donation join the DSR and match with other families who used the same donor ID at the same sperm bank. The resulting donor pedigree consists of heterosexual, lesbian, and single mother families who are connected through the common anonymous sperm donor used to conceive their children. Here, we introduce a new quantitative genetic study design based on theunprecedented family relationshipsfound in the donor pedigree. We surveyed 945 individual families constituting 159 donor pedigrees from the Donor Sibling Registry and used their demographic, physical, and behavioral characteristics to conduct a quantitative genetic study of selection and heritability. A direct measurement of phenotypic assortment showed mothers actively selected mates for height, eye color, and religion. vArtificial selection for donor height increased mean child height in a manner consistent with the selection differential. Reared-apart donor-conceived paternal half-siblings providedunbiased heritability estimates for traits influenced by maternal and contrast effects. Maternal effects were important in determining the variance of birth weight while eliminating contrast effects revealed sociability to be a highly heritable childhood temperament. Thus, the unprecedented family relationships in the donor pedigree enable a universal model for quantitative genetics.
[Arslan & Penke comment: "Estimates of heritability derived from twin studies have held up remarkably well when re-examined using different family relationships (e.g. parents, siblings, half- and adopted siblings) and can be easily extended to novel data such as the sometimes numerous offspring of sperm donors. In cases where selection is fairly clear-cut, estimates of heritability have borne out their usefulness as predictors of the response to selection. For example, children of sperm donors are taller in a manner consistent with their mothers’ selection on donor height (J.C. Lee, 2013)."]
[My comment: I can't believe I've never seen a study like this before. Sperm/egg donation is a great adjunct to classic twin and adoption studies. That it lets one verify that selection+heritability has the eugenic effects expected is even better.]
1.1.3 Specialty Designs: Specialty study designs explore a wide range of family relationships:
- Twins reared apart – Studies of twins reared apart fuse twin and adoption methodologies to eliminate shared environment. Twins separated at birth are extremely rare (9, 10).
- Children of MZ twins – Nominally cousins, children of MZ twins are genetic half-siblings who are reared apart in different households (11, 12).
- Half-siblings – In an indirect study, half-siblings are compared with full-siblings to eliminate the effect of shared environment (13-17). The more powerful direct study examines half-siblings who have been reared apart. Such half-siblings may vary in their exposure to a common environment or biological parent (8).
- Exact genetic relationship – Genome-wide DNA markers can be used to determine the exact coefficient of relatedness. The variation in r between full siblings enables a direct estimate of the heritability (18). This approach can be taken to its logical extreme using SNPs to calculate distant genetic relationships between putatively unrelated individuals, but this method cannot yet fully account for observed heritability (19).
The theory of quantitative genetics has been thoroughly tested and applied in plants and animals, but the genetic architecture of complex traits and the response to selection are open questions in humans. Here, we introduce a new family study design to address these issues.
The donor pedigree is a historically unprecedented family structure made possible by modern reproductive medicine. It consists of heterosexual, lesbian, and single mother families who are connected through the common anonymous sperm donor used to conceive their children. We used the unique gender and kinship arrangements found in the donor pedigree to conduct a quantitative genetic study with three aims: (i) to examine female mate choice preferences and determine which traits undergo active selection in humans, (ii) to describe the response to selection as a realized heritability, and (iii) to establish donor-conceived reared-apart paternal half-siblings as a model to measure the heritability of traits intractable to other study designs.
Study data was collected and managed using REDCap electronic data capture tools hosted at UCSF (12). Subjects self-identified as either a biological mother of a donorconceived child, a donor, a donor-conceived child (age 7-12), a donor-conceived adolescent (age 13-18), a donor-conceived adult (age 18+), or a non-biological parent of a donor-conceived child. Subjects completed surveys tailored to their self-identified group.
We restricted our analysis to data from biological mother reports due to the low absolute number of self-reports from donors, donor-conceived children, donor-conceived adolescents, donor-conceived adults, and non-biological parents. We further restricted our analysis to complete paternal half-sibling pedigrees that included a biological mother, a sperm donor with a known ID and clinic, donor-conceived children, and a nonbiological parent (if applicable). The complete pedigree requirement excludes mothers who used egg donation, mothers who did not have full donor ID/clinic information, and mothers who did not advance far enough into the survey to complete the parent report on their child.
Biological mothers completed self-report surveys regarding demographic, physical, and behavioral characteristics. They provided information about their donor and partner (if applicable) and then completed a parent report about their donor-conceived child's physical characteristics, medical history, temperament, symptoms of mental disorders, and birth and early development. Temperament was evaluated using the Emotionality, Activity, and Sociability (EAS) Temperament Survey for Children (13). Symptoms of mental disorders were measured using the Strengths and Difficulties Questionnaire (SDQ) (14). Birth and early development events were assessed using the NCS-A birth and early development questionnaire (15). Variable definitions for survey responses are shown in Table 2.1.
Parent age, height, and body mass index (BMI) distributions are shown in Figure 2.2. Male partners displayed greater positive skew for age than mothers or female partners. Donors were taller and had lower BMIs than male partners. The complete pedigrees contain a total of 1213 children. Descriptive statistics for these children are shown in Table 2.4. The prevalence of multiple births in the DSR was higher than the US national average of 3.3% due to the use of assisted reproductive technology (16).
Child age, height, and BMI distributions are shown in Figure 2.3. Donor-conceived children were taller and had higher BMIs compared to median CDC growth curves (17). Cross-tabulation tables for mother/donor/child eye and hair color are shown in Table 2.5 and Table 2.6, respectively. These tables are symmetric across the main diagonal, demonstrating internal consistency with regards to eye and hair color genetics. Table 2.7 shows the distribution of the number of children per biological mother. Approximately 25% of mothers had more than one donor-conceived child, either as a result of multiple births or from multiple donor-assisted conceptions. Based on shared donor ID and clinic information, 576 out of the 1213 children matched with a half-sibling internally within our sample, capturing 8.1% of the 7155 children who had matched with a half-sibling in the DSR at the start of the study. The 576 children who matched with a half-sibling formed 159 donor pedigrees in which each child shares the same sperm donor. The distribution of the number of children in each donor pedigree is shown in Table 2.8; the largest donor pedigree contains 10 children, with a median size of three children.
We assessed the reliability of the donor ID/clinic half-sibling matching process by measuring the inter-rater reliability of mother-reported physical characteristics for the donors. Krippendorff alpha (18) values for height ( =0.67), weight ( =0.76), eye color ( =0.78), and hair color ( = 0.79) support the claim that mothers in each donor pedigree used the same donor.
Figure 2.5 shows maternal age at time of birth. Although maternal age was not a significant covariate for birth weight, the spike at age 38 illustrates the inevitability of the biological clock.
Numerous surveys have been conducted to determine which traits women value when selecting a mate, but they all suffer from the same fault: women's stated preferences may not be reflected in their mate choices (2-5).
Anonymous sperm donation eliminates these confounding processes to reveal a clear link between female mate choice preferences and mate selection. Mothers freely choose a donor from a sperm bank, particularly single mothers who express their preferences without input from a partner. Thus, the single mother/donor correlation is the first direct measure of phenotypic assortment in humans, standing in contrast to previous indirect methods based on twins and their spouses (5, 11-13). Any heterosexual mother/male partner or lesbian mother/female partner correlation in excess of the single mother/donor correlation is attributable to social homogamy (Table 2.2).
The majority of mothers said race/ethnicity, education level, height, body mass index (BMI), eye color, and hair color were important, but employment status and religion were not (Table 3.3). Given the donor's wholly genetic contribution, these ratings can be interpreted as reflecting mothers' beliefs about heritability.
[Silly mothers. Of course employment status and religion are heritable. Employment, if nothing else, through race, education, and intelligence.]
3.3.2.1 Height and Eye Color
We used phenotypic assortment to determine which traits underwent active selection. For height and eye color, the single mother/donor correlations were greater than the heterosexual mother/male partner and lesbian mother/female partner correlations (Table 3.3). This suggests mutual mate choice limits phenotypic assortment in a monogamous mating system, as compared to unconstrained female choice of donor (15). A log-linear regression analysis of eye color shows single mothers preferentially selected donors with recessive blue eyes (Table 3.4, Figure 3.1). Thus, mothers actively selected for height and eye color in accordance with their stated preferences.
Examining the remaining parental correlations for height and eye color, the positive male partner/donor and female partner/donor correlations (Table 3.3) indicate biological mothers in heterosexual and lesbian couples chose a donor to match their partner in a transitive form of phenotypic assortment. Log-linear regression results show heterosexual and lesbian couples matched partner and donor eye colors (Table 3.4, Figure 3.2, Figure 3.3). The lesbian mother/donor correlations were significant, but the heterosexual mother/donor correlations were not. Lesbian couples could choose a donor to match either parent because they do not contend with the same paternity issues facing heterosexual couples (16, 17).
3.3.2.2 Education Level and BMI
For education level and BMI, the heterosexual mother/male partner (education level only) and lesbian mother/female partner correlations were significant, but the single mother/donor correlations were not (Table 3.3). Assortative mating for education level 47and BMI was therefore driven by passive social homogamy, contrary to expectations from mothers' stated preferences. The single mother/donor correlations may have been influenced by a ceiling effect in which the majority of donors were college-educated and had healthy BMIs (Table 2.3).
3.3.2.3 Religion
For religion, the heterosexual mother/male partner and lesbian mother/female partner correlations were greater than the single mother/donor correlation (Table 3.3). All three correlations were significant, signaling the influence of both phenotypic assortment and social homogamy. Phenotypic assortment was driven by the association of atheist/atheist and Jewish/Jewish pairings between single mothers and donors (Table 3.5, Figure 3.4). Social homogamy was driven by atheist/atheist, Jewish/Jewish, and other/other pairings between mothers and their partners (Table 3.5, Figure 3.5, Figure 3.6). Judaism's dual role as a religion and an ethnicity could partially explain why there was phenotypic assortment for religion despite it not being an important factor in donor choice.
3.3.2.4 Race/Ethnicity
For race/ethnicity, our sample was approximately 90% white (Table 2.3). Heterosexual couples (Figure 3.7) and single mothers (Figure 3.9) were less diverse then lesbian couples (Figure 3.8). Lesbian couples demonstrated concordance for race/ethnicity via statistically significant correlations (Table 3.3) and associations (Table 3.6) between lesbian mother/female partner, lesbian mother/donor, and female partner/donor. The single mother/donor correlation was statistically significant (Table 3.3), but no log-linear regression associations were found between single mother and 48donor due to small non-white sample size (Table 3.6). The heterosexual mother/male partner and heterosexual mother/donor correlations were not significant because the small number of non-white mothers in heterosexual couples almost exclusively had white male partners and they all chose white donors (Figure 3.7). This idiosyncratic pattern and lack of diversity precludes a general inference about selection for race/ethnicity.
3.3.2.5 Hair Color, and Employment Status
We did not observe any assortative mating for hair color or employment status between single mother/donor, heterosexual mother/male partner, or lesbian mother/female partner. Lesbian couples matched partner/donor hair color, while heterosexual couples did not (Table 3.3).
Children in the DSR were taller than the median growth curve by R = 1.23 inches, averaged across all ages for both sexes (Figure 2.3B). The selection differential, S , measures the strength of selection and is defined as the difference between the mean height of the selected parents and the mean height of the population. Biological mothers were taller than the median Caucasian female by 0.7 inches and selected donors were taller than the median Caucasian male by 2 inches, resulting in a selection differential of S = 1.35 inches (Table 2.3) (18). The response to selection is related to the selection differential by the realized heritability h^2 = R / S . Assisted reproduction created a rare natural experiment to study artificial selection for height in humans; the effect of selection as described by the realized heritability h^2 = 0 .91 was consistent with the heritability of adult height calculated using traditional methods (19).
The donor pedigree enables the first direct measure of phenotypic assortment in humans. We compared mother/donor with mother/partner correlations and found mothers actively selected for height, eye color, and religion. The response to selection for height matched theoretical predictions; taller donors begat taller children in a manner consistent with previous heritability estimates. Our results represent a unique experimental validation of ethical artificial selection in humans. Mothers who selected for height endowed their children with an economic advantage because height is positively associated with social status and labor market outcomes (20)."
"Quantitative Genetics in the Postmodern Family of the Donor Sibling Registry", Lee 2013:
"Quantitative genetics is primarily concerned with two subjects: the correlation between relatives and the response to selection. The correlation between relatives is used to determinethe heritability of a trait--the key quantity that addresses the question of nature vs. nurture. Heritability, in turn, is used to predict the response to selection --the maindriver ofimprovementsin crops and livestock. The theory of quantitative genetics has been thoroughly tested and applied in plants and animals, but heritability and selection remain open questions in humans due to limited natural experimental designs. The Donor Sibling Registry (DSR)is an organization that helps individuals conceived as a result of sperm, egg, or embryo donation make contact withgenetically related individuals. Families who conceived children via anonymous sperm donation join the DSR and match with other families who used the same donor ID at the same sperm bank. The resulting donor pedigree consists of heterosexual, lesbian, and single mother families who are connected through the common anonymous sperm donor used to conceive their children. Here, we introduce a new quantitative genetic study design based on theunprecedented family relationshipsfound in the donor pedigree. We surveyed 945 individual families constituting 159 donor pedigrees from the Donor Sibling Registry and used their demographic, physical, and behavioral characteristics to conduct a quantitative genetic study of selection and heritability. A direct measurement of phenotypic assortment showed mothers actively selected mates for height, eye color, and religion. vArtificial selection for donor height increased mean child height in a manner consistent with the selection differential. Reared-apart donor-conceived paternal half-siblings providedunbiased heritability estimates for traits influenced by maternal and contrast effects. Maternal effects were important in determining the variance of birth weight while eliminating contrast effects revealed sociability to be a highly heritable childhood temperament. Thus, the unprecedented family relationships in the donor pedigree enable a universal model for quantitative genetics.
[Arslan & Penke comment: "Estimates of heritability derived from twin studies have held up remarkably well when re-examined using different family relationships (e.g. parents, siblings, half- and adopted siblings) and can be easily extended to novel data such as the sometimes numerous offspring of sperm donors. In cases where selection is fairly clear-cut, estimates of heritability have borne out their usefulness as predictors of the response to selection. For example, children of sperm donors are taller in a manner consistent with their mothers’ selection on donor height (J.C. Lee, 2013)."]
[My comment: I can't believe I've never seen a study like this before. Sperm/egg donation is a great adjunct to classic twin and adoption studies. That it lets one verify that selection+heritability has the eugenic effects expected is even better.]
1.1.3 Specialty Designs: Specialty study designs explore a wide range of family relationships:
- Twins reared apart – Studies of twins reared apart fuse twin and adoption methodologies to eliminate shared environment. Twins separated at birth are extremely rare (9, 10).
- Children of MZ twins – Nominally cousins, children of MZ twins are genetic half-siblings who are reared apart in different households (11, 12).
- Half-siblings – In an indirect study, half-siblings are compared with full-siblings to eliminate the effect of shared environment (13-17). The more powerful direct study examines half-siblings who have been reared apart. Such half-siblings may vary in their exposure to a common environment or biological parent (8).
- Exact genetic relationship – Genome-wide DNA markers can be used to determine the exact coefficient of relatedness. The variation in r between full siblings enables a direct estimate of the heritability (18). This approach can be taken to its logical extreme using SNPs to calculate distant genetic relationships between putatively unrelated individuals, but this method cannot yet fully account for observed heritability (19).
The theory of quantitative genetics has been thoroughly tested and applied in plants and animals, but the genetic architecture of complex traits and the response to selection are open questions in humans. Here, we introduce a new family study design to address these issues.
The donor pedigree is a historically unprecedented family structure made possible by modern reproductive medicine. It consists of heterosexual, lesbian, and single mother families who are connected through the common anonymous sperm donor used to conceive their children. We used the unique gender and kinship arrangements found in the donor pedigree to conduct a quantitative genetic study with three aims: (i) to examine female mate choice preferences and determine which traits undergo active selection in humans, (ii) to describe the response to selection as a realized heritability, and (iii) to establish donor-conceived reared-apart paternal half-siblings as a model to measure the heritability of traits intractable to other study designs.
Study data was collected and managed using REDCap electronic data capture tools hosted at UCSF (12). Subjects self-identified as either a biological mother of a donorconceived child, a donor, a donor-conceived child (age 7-12), a donor-conceived adolescent (age 13-18), a donor-conceived adult (age 18+), or a non-biological parent of a donor-conceived child. Subjects completed surveys tailored to their self-identified group.
We restricted our analysis to data from biological mother reports due to the low absolute number of self-reports from donors, donor-conceived children, donor-conceived adolescents, donor-conceived adults, and non-biological parents. We further restricted our analysis to complete paternal half-sibling pedigrees that included a biological mother, a sperm donor with a known ID and clinic, donor-conceived children, and a nonbiological parent (if applicable). The complete pedigree requirement excludes mothers who used egg donation, mothers who did not have full donor ID/clinic information, and mothers who did not advance far enough into the survey to complete the parent report on their child.
Biological mothers completed self-report surveys regarding demographic, physical, and behavioral characteristics. They provided information about their donor and partner (if applicable) and then completed a parent report about their donor-conceived child's physical characteristics, medical history, temperament, symptoms of mental disorders, and birth and early development. Temperament was evaluated using the Emotionality, Activity, and Sociability (EAS) Temperament Survey for Children (13). Symptoms of mental disorders were measured using the Strengths and Difficulties Questionnaire (SDQ) (14). Birth and early development events were assessed using the NCS-A birth and early development questionnaire (15). Variable definitions for survey responses are shown in Table 2.1.
Parent age, height, and body mass index (BMI) distributions are shown in Figure 2.2. Male partners displayed greater positive skew for age than mothers or female partners. Donors were taller and had lower BMIs than male partners. The complete pedigrees contain a total of 1213 children. Descriptive statistics for these children are shown in Table 2.4. The prevalence of multiple births in the DSR was higher than the US national average of 3.3% due to the use of assisted reproductive technology (16).
Child age, height, and BMI distributions are shown in Figure 2.3. Donor-conceived children were taller and had higher BMIs compared to median CDC growth curves (17). Cross-tabulation tables for mother/donor/child eye and hair color are shown in Table 2.5 and Table 2.6, respectively. These tables are symmetric across the main diagonal, demonstrating internal consistency with regards to eye and hair color genetics. Table 2.7 shows the distribution of the number of children per biological mother. Approximately 25% of mothers had more than one donor-conceived child, either as a result of multiple births or from multiple donor-assisted conceptions. Based on shared donor ID and clinic information, 576 out of the 1213 children matched with a half-sibling internally within our sample, capturing 8.1% of the 7155 children who had matched with a half-sibling in the DSR at the start of the study. The 576 children who matched with a half-sibling formed 159 donor pedigrees in which each child shares the same sperm donor. The distribution of the number of children in each donor pedigree is shown in Table 2.8; the largest donor pedigree contains 10 children, with a median size of three children.
We assessed the reliability of the donor ID/clinic half-sibling matching process by measuring the inter-rater reliability of mother-reported physical characteristics for the donors. Krippendorff alpha (18) values for height ( =0.67), weight ( =0.76), eye color ( =0.78), and hair color ( = 0.79) support the claim that mothers in each donor pedigree used the same donor.
Figure 2.5 shows maternal age at time of birth. Although maternal age was not a significant covariate for birth weight, the spike at age 38 illustrates the inevitability of the biological clock.
Numerous surveys have been conducted to determine which traits women value when selecting a mate, but they all suffer from the same fault: women's stated preferences may not be reflected in their mate choices (2-5).
Anonymous sperm donation eliminates these confounding processes to reveal a clear link between female mate choice preferences and mate selection. Mothers freely choose a donor from a sperm bank, particularly single mothers who express their preferences without input from a partner. Thus, the single mother/donor correlation is the first direct measure of phenotypic assortment in humans, standing in contrast to previous indirect methods based on twins and their spouses (5, 11-13). Any heterosexual mother/male partner or lesbian mother/female partner correlation in excess of the single mother/donor correlation is attributable to social homogamy (Table 2.2).
The majority of mothers said race/ethnicity, education level, height, body mass index (BMI), eye color, and hair color were important, but employment status and religion were not (Table 3.3). Given the donor's wholly genetic contribution, these ratings can be interpreted as reflecting mothers' beliefs about heritability.
[Silly mothers. Of course employment status and religion are heritable. Employment, if nothing else, through race, education, and intelligence.]
3.3.2.1 Height and Eye Color
We used phenotypic assortment to determine which traits underwent active selection. For height and eye color, the single mother/donor correlations were greater than the heterosexual mother/male partner and lesbian mother/female partner correlations (Table 3.3). This suggests mutual mate choice limits phenotypic assortment in a monogamous mating system, as compared to unconstrained female choice of donor (15). A log-linear regression analysis of eye color shows single mothers preferentially selected donors with recessive blue eyes (Table 3.4, Figure 3.1). Thus, mothers actively selected for height and eye color in accordance with their stated preferences.
Examining the remaining parental correlations for height and eye color, the positive male partner/donor and female partner/donor correlations (Table 3.3) indicate biological mothers in heterosexual and lesbian couples chose a donor to match their partner in a transitive form of phenotypic assortment. Log-linear regression results show heterosexual and lesbian couples matched partner and donor eye colors (Table 3.4, Figure 3.2, Figure 3.3). The lesbian mother/donor correlations were significant, but the heterosexual mother/donor correlations were not. Lesbian couples could choose a donor to match either parent because they do not contend with the same paternity issues facing heterosexual couples (16, 17).
3.3.2.2 Education Level and BMI
For education level and BMI, the heterosexual mother/male partner (education level only) and lesbian mother/female partner correlations were significant, but the single mother/donor correlations were not (Table 3.3). Assortative mating for education level 47and BMI was therefore driven by passive social homogamy, contrary to expectations from mothers' stated preferences. The single mother/donor correlations may have been influenced by a ceiling effect in which the majority of donors were college-educated and had healthy BMIs (Table 2.3).
3.3.2.3 Religion
For religion, the heterosexual mother/male partner and lesbian mother/female partner correlations were greater than the single mother/donor correlation (Table 3.3). All three correlations were significant, signaling the influence of both phenotypic assortment and social homogamy. Phenotypic assortment was driven by the association of atheist/atheist and Jewish/Jewish pairings between single mothers and donors (Table 3.5, Figure 3.4). Social homogamy was driven by atheist/atheist, Jewish/Jewish, and other/other pairings between mothers and their partners (Table 3.5, Figure 3.5, Figure 3.6). Judaism's dual role as a religion and an ethnicity could partially explain why there was phenotypic assortment for religion despite it not being an important factor in donor choice.
3.3.2.4 Race/Ethnicity
For race/ethnicity, our sample was approximately 90% white (Table 2.3). Heterosexual couples (Figure 3.7) and single mothers (Figure 3.9) were less diverse then lesbian couples (Figure 3.8). Lesbian couples demonstrated concordance for race/ethnicity via statistically significant correlations (Table 3.3) and associations (Table 3.6) between lesbian mother/female partner, lesbian mother/donor, and female partner/donor. The single mother/donor correlation was statistically significant (Table 3.3), but no log-linear regression associations were found between single mother and 48donor due to small non-white sample size (Table 3.6). The heterosexual mother/male partner and heterosexual mother/donor correlations were not significant because the small number of non-white mothers in heterosexual couples almost exclusively had white male partners and they all chose white donors (Figure 3.7). This idiosyncratic pattern and lack of diversity precludes a general inference about selection for race/ethnicity.
3.3.2.5 Hair Color, and Employment Status
We did not observe any assortative mating for hair color or employment status between single mother/donor, heterosexual mother/male partner, or lesbian mother/female partner. Lesbian couples matched partner/donor hair color, while heterosexual couples did not (Table 3.3).
Children in the DSR were taller than the median growth curve by R = 1.23 inches, averaged across all ages for both sexes (Figure 2.3B). The selection differential, S , measures the strength of selection and is defined as the difference between the mean height of the selected parents and the mean height of the population. Biological mothers were taller than the median Caucasian female by 0.7 inches and selected donors were taller than the median Caucasian male by 2 inches, resulting in a selection differential of S = 1.35 inches (Table 2.3) (18). The response to selection is related to the selection differential by the realized heritability h^2 = R / S . Assisted reproduction created a rare natural experiment to study artificial selection for height in humans; the effect of selection as described by the realized heritability h^2 = 0 .91 was consistent with the heritability of adult height calculated using traditional methods (19).
The donor pedigree enables the first direct measure of phenotypic assortment in humans. We compared mother/donor with mother/partner correlations and found mothers actively selected for height, eye color, and religion. The response to selection for height matched theoretical predictions; taller donors begat taller children in a manner consistent with previous heritability estimates. Our results represent a unique experimental validation of ethical artificial selection in humans. Mothers who selected for height endowed their children with an economic advantage because height is positively associated with social status and labor market outcomes (20)."
Shared publicly