FADS1and the timing of human adaptation to agriculture

Abstract

1Variation at the FADS1/FADS2 gene cluster is functionally associated with differences in lipid 2 metabolism and is often hypothesized to reflect adaptation to an agricultural diet. Here, we test the 3 evidence for this relationship using both modern and ancient DNA data. We show that almost all the 4 inhabitants of Europe carried the ancestral allele until the derived allele was introduced approximately 5 8,500 years ago by Early Neolithic farming populations. However, we also show that it was not under 6 strong selection in these populations. We find that this allele, and other proposed agricultural adaptations 7 at LCT/MCM6 and SLC22A4, were not strongly selected until much later, perhaps as late as the Bronze 8 Age. Similarly, increased copy number variation at the salivary amylase gene AMY1 is not linked to the 9 development of agriculture although, in this case, the putative adaptation precedes the agricultural 10 transition. Our analysis shows that selection at the FADS locus was not tightly linked to the initial 11 introduction of agriculture and the Neolithic transition. Further, it suggests that the strongest signals of 12 recent human adaptation in Europe did not coincide with the Neolithic transition but with more recent 13 changes in environment, diet or efficiency of selection due to increases in effective population size. 56Greenlandic Inuit population reduces the activity of FADS1, is associated with PUFA levels, and is likely 57 an adaptation to a diet that is extremely high in LC-PUFA from marine sources (Fumagalli, et al. 2015). 58Conversely, a variant that increases expression of FADS2 has been selected in South Asian populations -59 and may be a specific adaptation to a vegetarian diet (Kothapalli, et al. 2016). Reflecting its important 60 role in lipid metabolism, variation at the FADS locus also contributes significantly to variation in lipid 61 levels in present-day populations. As well as directly contributing to variation in PUFA levels, SNPs in 62 haplotype D are among the strongest genome-wide association study (GWAS) signals for triglyceride and 63 cholesterol levels (Teslovich, et al. 2010). Thus, the complex evolutionary history of the region is not 64 only informative about ancient human diets, but also potentially relevant for understanding the 65 distribution of lipid-related disease risk both within and between populations. We therefore aimed to 66 3 characterize the history of the region, by combining inference from ancient and modern DNA data, in 67 order to understand the evolutionary basis of this important functional variation and its relationship with 68 changes in diet. We also compared the evolutionary history of the FADS locus with the histories of other 69 loci involved in dietary adaptation, to see whether we could detect shared patterns of adaptation. 70

Genetic contributions to variation in human stature in prehistoric Europe

“…In particular, the Mesolithic/Neolithic overlap was resolved based on whether each individual had more (Neolithic) or less (Mesolithic) than 50% ancestry related to northwest Anatolian Neolithic Farmers. The Neolithic/post-Neolithic overlap was resolved based on whether individuals had more than 25% ancestry related to Bronze Age Steppe populations (“Steppe ancestry”; See (83) for details). For the skeletal data, group assignment in the overlapping periods was determined by the archaeology of each site.…”

Genome-wide analysis identifies genetic effects on reproductive success and ongoing natural selection at theFADSlocus

“…In particular, the derived haplotype at this locus has increased from a frequency of <10% 10,000 years ago to 60-75% in present-day European populations ( Figure 3B). While some of this increase is due to admixture, there is strong evidence of positive selection over the past few thousand years, even accounting for changes in ancestry 17,[22][23][24][25][26] , Each 'C' allele of the lead NEB SNP rs108499, which tags the positively selected FADS1 haplotype, increased NEB by 0.0134, corresponding to a selection coefficient of 0.74% (0.0134 divided by mean NEB of 1.8).…”

“…Consistent with this, we estimate that within the "White British" subset of UK Biobank, the derived allele increased in frequency by 0.0088% per-year between the 1938 and 1969 birth cohorts (linear regression including 10 principal components and collection centre; P=0.27, Figure 3C) corresponding to a selection coefficient of 1.2% (approximate 95% CI -0.9-3.2%). Estimates of historical selection coefficients range from 0.4-0.6% (based on time-serial analysis of ancient DNA samples) to 3.4-6.9% (based on analyses of present-day haplotype structure) 24,26 .…”

Limited evidence for selection at theFADSlocus in Native American populations

“…Haplotype D appears to have been under selection-likely preceding the 38 out-of-Africa bottleneck-in Africa, and is virtually fixed in present-day African populations 39 (Ameur, et al 2012;Mathias, et al 2012). Given this, it is surprising that early Eurasian 40 populations appear to have largely carried the ancestral haplotype, suggesting selection for the 41 ancestral haplotype at some time after the split of present-day African and non-African lineages 42 (Ye, et al 2017;Mathieson and Mathieson 2018). By the Mesolithic-around 10,000 years 43 before present (BP)-the ancestral haplotype was fixed in Europe (Mathieson, et al 2015).…”

“…For each PEL individual, 87we restricted to regions of homozygous Native American ancestry(Martin, et al 2017). We 88 used CEU (Northern and Western European ancestry) to represent present-day Europeans, CHB 89 (Chinese from Beijing) to represent East Asians and 118 ancient European hunter-gatherers 90 (HG) to represent Mesolithic Europe(Mathieson and Mathieson 2018). We replicate the 91 elevated PBS(PEL, (CEU, CHB)) statistic at the FADS locus (Fig.2 left column; upper 0.0002 92 quantile), but find that it largely disappears if we replace CHB with HG (Fig.…”