One hundred years ago, Fisher (1918) published a paper that highlighted that genetic researchers should be especially interested in cousin pairs. We have the potential using the NLSY data to respond to a 100-year-old challenge issued by Fisher: ‘‘… the hypothesis of cumulative Mendelian factors seems to fit the facts very accurately. The only marked discrepancy from existing published work lies in the correlation for first cousins. … but until we have a record of complete cousinships measured accurately and without selection, it will not be possible to obtain satisfactory numerical evidence on this question’’ (p. 168). Like Fisher, we have observed the “cousin anomaly” in the NLSY cousin data. Further, the NLSY data approximate Fisher’s “complete cousinships measured accurately and without selection.”
In the current study, we compute kinship correlations, and estimate ACE models, to directly address Fisher’s question, and also indirectly expand it to include several different questions:
- Are cousins anomalous in the kinship relatedness on a highly reliable phenotype, height (Fisher’s challenge)?
- What is the biometrical status of other cousin categories besides full cousins?
- Can we conduct a genetically-informed biometrical study with sufficient power using only a cousin categories?
- If so, what results do we obtain?
Recently developed kinship links in the NLSY79 and NLSY-Children have produced thousands of cousin pairs. A few cousin studies exist in the literature, but none (to our knowledge) have used biometric modeling of only cousin categories. Using cousins is challenging, partly because different cousin categories are so similar in their average (segregating) genetic relatedness; large samples are required to have enough statistical power. Alternatively, a highly reliable phenotype can also contribute. We use height as the phenotype in the current study, along with an impressively large sample of cousins. Table 4 in Rodgers et al (2016) lists and describes nine cousin categories in the NLSY-Children data.
We used full cousins (R=.125, N=4179) and half cousins (R=.0625, N=548), and estimated an ACE model of self-reported height (after age 17, standardized within gender); we set the C parameter to .80 (compared to the usual 1.0 for those raised in the same house). Kinship correlations were r=.17 for full cousins, and r=.10 for half cousins. The ACE estimates were a2 = .92, c2 = .07, e2 = .01; these are very close to traditional estimates of height heritability and c2.
We will engage additional biometrical analyses to investigate NLSY cousin pairs, partly for substantive interest, but also as proof-in-principle of the value of the NLSY cousin data, and to help address Fisher’s challenge.
References
Fisher RA (1918) The correlation between relatives on the supposition of Mendelian inheritance. Philos Trans R Soc Edinburgh 52:399–433
Rodgers et al (2016 ). The NLSY Kinship Links: Using the NLSY79 and NLSY-Children Data to Conduct Genetically-Informed and Family-Oriented Research. Behavior Genetics.
