Associations between human genetic and craniometric differentiation across North Eurasia: The role of geographic scale

Abstract

This study sets out to consider the influence of geographical scale on the association between molecular genetic differentiation and craniometric phenotypic differentiation in recent human populations. We employ interpopulation distance measurements for three different anatomical regions of the skull and for three different systems of genetic markers in 30 Eurasian populations. Our original dataset comprises 703 male skulls measured for 21 mid-facial, 15 neurocranial and 6 mandibular measurements, in all cases assessing Mahalanobis distances between populations. Published genetic data of more than 2,000 individuals were summarized by between-population FST based on allele frequencies of autosomal single nucleotide polymorphisms (SNPs), as well as Cavalli-Sforza distances based on the frequencies of 19 Y-chromosome and 29 mtDNA haplogroups. For different geographical scales of analysis, we used Mantel tests to assess the association of craniometric and genetic inter-population distances for the different cranial regions and genetic markers. Our results show that the level of association between craniometric and genetic distances depends on the part of the skull quantified and on the set of variables employed. In our dataset, this association is much stronger for the mid-face than for the cranial vault. Furthermore, the Mantel test correlation coefficients for the broadest, intercontinental level of analysis are moderate to high, and some are among the highest published so far. They are consistently lower at smaller geographic levels of comparison. Autosomal SNP distances exhibit the strongest associations with cranial morphology for almost all anatomical regions and at all geographical levels. Our results are evaluated against the background of previous studies assessing the correlation between craniometric, genetic, and geographic distances, drawing attention to the need for investing much more effort in studying factors affecting the association between genetic and craniometric distances at regional and local geographical levels.

Inferring human neutral genetic variation from craniodental phenotypes

“…A standard approach for quantifying the utility of a given craniodental data type in capturing a neutral genomic signature is to estimate phenotypic distances among worldwide modern human populations, on the one hand, and to compare them to neutral genomic distances estimated among the same or closely matched set of populations on the other ( 1 , 2 , 52 ). These analyses, hereafter termed D P – D G comparisons, have been extensively performed for cranial metric data ( 14 , 16 , 17 , 19 , 20 , 51 , 53 , 54 ), dental metric data ( 18 , 55 ), cranial nonmetric trait data ( 51 , 56 , 57 ), and dental nonmetric trait data ( 10 , 15 , 18 , 58 ). However, the estimated levels of neutrality of the different craniodental data types reported in previous D P – D G studies are not directly comparable, since different populations have been sampled and diverse methodological approaches for calculating between-population distances have been employed at different geospatial scales ( 54 ).…”

“…These analyses, hereafter termed D P – D G comparisons, have been extensively performed for cranial metric data ( 14 , 16 , 17 , 19 , 20 , 51 , 53 , 54 ), dental metric data ( 18 , 55 ), cranial nonmetric trait data ( 51 , 56 , 57 ), and dental nonmetric trait data ( 10 , 15 , 18 , 58 ). However, the estimated levels of neutrality of the different craniodental data types reported in previous D P – D G studies are not directly comparable, since different populations have been sampled and diverse methodological approaches for calculating between-population distances have been employed at different geospatial scales ( 54 ).…”