Mixed Race Studies

Examining Population Stratification via Individual Ancestry Estimates versus Self-Reported Race

Cancer Epidemiology, Biomarkers & Prevention
Volume 14, Issue 6 (June 2005)
pages 1545-1551
DOI: 10.1158/1055-9965.EPI-04-0832

Jill S. Barnholtz-Sloan
Cancer Prevention and Control Program
H. Lee Moffitt Cancer Center and Research Institute

Ranajit Chakraborty
Center for Genome Research, Department of Environmental Health
University of Cincinnati, Cincinnati, Ohio

Thomas A. Sellers
Cancer Prevention and Control Program
H. Lee Moffitt Cancer Center and Research Institute

Ann G. Schwartz
Population Studies and Prevention Program, Karmanos Cancer Institute and Department of Internal Medicine
Wayne State University School of Medicine, Detroit, Michigan

Population stratification has the potential to affect the results of genetic marker studies. Estimating individual ancestry provides a continuous measure to assess population structure in case-control studies of complex disease, instead of using self-reported racial groups. We estimate individual ancestry using the Federal Bureau of Investigation CODIS Core short tandem repeat set of 13 loci using two different analysis methods in a case-control study of early-onset lung cancer. Individual ancestry proportions were estimated for “European” and “West African” groups using published allele frequencies. The majority of Caucasian, non-Hispanics had >50% European ancestry, whereas the majority of African Americans had <20% European ancestry, regardless of ancestry estimation method, although significant overlap by self-reported race and ancestry also existed. When we further investigated the effect of ancestry and self-reported race on the frequency of a lung cancer risk genotype, we found that the frequency of the GSTM1 null genotype varies by individual European ancestry and case-control status within self-reported race (particularly for African Americans). Genetic risk models showed that adjusting for individual European ancestry provided a better fit to the data compared with the model with no group adjustment or adjustment for self-reported race. This study suggests that significant population substructure differences exist that self-reported race alone does not capture and that individual ancestry may be confounded with disease status and/or a candidate gene risk genotype.

Read the entire article here.

Gene admixture in human populations: Models and predictions

American Journal of Physical Anthropology
Volume 29, Issue Supplement S7 (1986)
pages 1–43
DOI: 10.1002/ajpa.1330290502

Ranajit Chakraborty, Robert A. Kehoe Professor and Director of Center for Genome Information
University of Cincinnati

Brief accounts of methods for estimating proportions of admixture in populations and individuals of hybrid origin are presented with the objective of appraising their underlying assumptions. In view of the uncertainties introduced by assumptions under which admixture estimates are obtained, it is concluded that the reliability of estimates derived from different methods cannot be formally compared. With examples from several admixed populations, it is shown that all methods do not necessarily give discordant results when identical data are used to obtain admixture estimates.

Even though past experiences using admixed populations to detect selection or to understand disease etiology have not been very successful, it is believed that admixed human populations can be regarded as a natural experiment. Hence, they are suitable for microevolutionary and epidemiological studies.

The proper identification of ancestral populations and the degree of asymmetry in gene flow (sex-biased admixture) are important issues in admixture studies. These aspects can be examined in the statistical properties of allele frequency distributions, but corroboration of particular models should be made from in-depth investigations of historical demography and social structure of admixed populations.

Future studies of admixture with DNA polymorphism data may resolve some of the uncertainties associated with current techniques of detecting genetic polymorphisms. Because of the abundance of genetic data, it is argued that morphological traits are of limited use in resolving current problems of human admixture studies.

Read or purchase the article here.