Social and genetic diversity in first farmers of central Europe

Abstract

The Linearbandkeramik (LBK) Neolithic communities were the first to spread farming across large parts of Europe. We report genome-wide data for 250 individuals: 178 individuals from whole-cemetery surveys of the Alföld Linearbankeramik Culture eastern LBK site of Polgár-Ferenci-hát, the western LBK site of Nitra Horné Krškany and the western LBK settlement and massacre site of Asparn-Schletz, as well as 48 LBK individuals from 16 other sites and 24 earlier Körös and Starčevo individuals from 17 more sites. Here we show a systematically higher percentage of western hunter-gatherer ancestry in eastern than in western LBK sites, showing that these two distinct LBK groups had different genetic trajectories. We find evidence for patrilocality, with more structure across sites in the male than in the female lines and a higher rate of within-site relatives for males. At Asparn-Schletz we find almost no relatives, showing that the massacred individuals were from a large population, not a small community.

Extended Figure 1:. Principal Component Analysis (PCA):

PCA performed with 879 modern Eurasian individuals in which the ancient individuals were projected. The modern individuals have been removed from the image. The PCA shows the clustering of the LBK and the position of individuals along the X axis, indicating differential WHG affinities and showing that WHG (represented by two Körös culture outliers with entirely WHG ancestry) are more closely related to ALPC. Three individuals: I6914 (Austria_LBK) and I1507, I497 (Köros) are outliers.

Extended Figure 2:. qpWave plots:

qpWave plots to test for individual differentiation, with each population represented in one plot. Grey colour means results were highly significant (little genetically related). The number after the name of each individual relates is the point estimate of WHG ancestry from qpAdm. A) ALPC individuals. Individuals I21898, I10349, I21902, I18660, I10350, I18656, I18695, I4186, I1499, I21714, and I2377 are labelled in our analysis as ALPC outliers with high WHG ancestry. Individuals: I21828, I21830, I10351, I10352, I10353, I18657, I21767, 17933, I1500, I2380, I3537, I17455, I18636, I29883, I18641 and I4187 are labelled in our analysis as ALPC outliers with low WHG ancestry. B) Austria LBK Individuals: Individuals I27785, I25349, I6913, I6912 and I24028 are labelled in our analysis as outliers with high WHG ancestry. C) Germany LBK Individuals, D) Slovakia LBK Individuals: Individual I18144 is labelled in our analysis as an outlier with high WHG ancestry. E) Transdanubia_Hungary LBK Individuals: individuals I1882 and I1883 are labelled in our analysis as outliers with high WHG ancestry. We used qpWave from admixtools to perform the plots, each square represents the two-sided p-value of every single test.

Extended Figure 3:. Parental haplogroups:

Distribution of the Y chromosome and mtDNA haplogroups per population. The Y-axis represents the number of individuals.

Extended Figure 4:. Isotopic data:

Isotope data from Pólgar-Ferenci-hát. Here we plot the ratio δ¹³C/δ¹⁵N. Each dot represents one individual and the colour denotes the family.

Extended Figure 5:. Rund of Homozigosity:

ROH distribution in the dataset. A)LBK individuals, B) ALPC individuals, C) Koros and Starcevo individuals. Individuals with more than 400,000 SNPs and the assessed ROH. Individuals in the ALPC group show a higher rate of close-kin unions (as reflected in the presence of ROH segments >20cM) than the rest of the dataset.

Extended Figure 6:. Natural selection in Neolithics:

Tests for positive selection in the ALPC and LBK population, made with qqman . The red lines indicate the top 0.05% cutoff. (A) Normalized iHS scores in ALPC. (B) Normalized iHS scores in LBK. (C) Normalized unphased iHS scores in ALPC. (D) Normalized unphased iHS scores in LBK. (E) Normalized nSL scores in ALPC. (F) Normalized nSL scores in LBK. (G) Normalized unphased nSL scores in ALPC. (H) Normalized unphased nSL scores in LBK.

Extended Figure 7:

Correspondence between the ancestry in ALPC and LBK segments with the selection scan values. Each dot represents a region of 0.2 cM of the genome, in the Y-axis we display the average WHG ancestry of the region, and in the X-axis the average selection scores from candidate SNPs within the region (Supplementary Table S11). We show the two-sided Spearman correlation coefficients and p-value.

Figure 1:. The LBK and ALPC extension:

A) Map of the extent of the LBK and ALPC cultures in Central Europe. Generated with Illustrator. The extent of the LBK and ALPC cultures was obtained from Gronenborn and Horejs 2023 . B) Location of the studied sites, the symbols depict the cultural attribution.

Figure 2:. The genomic ancestry diversity in the LBK/ALPC

A) Histograms of point estimates of ancestry proportions of LBK and ALPC individuals for both the autosomes and X-chromosome (generated with ggplot2 ). B) Range of dates and culture span of the individuals included in the study

Figure 3:. Kinship patterns in LBK sites:

Burial layouts for A) Nitra Horné Krškany (top) and B) Polgár-Ferenci-hát (bottom). Each symbol represents one individual: squares males, circles females. Red denotes the main genetic cluster, green WHG outliers, and violet EEF outliers. Light brown are children. Blue lines or circles are 1^st-degree relatives and the yellow pottery symbol grave goods in burials. Only individuals with ancestry information are plotted C) Dietary isotopes at Nitra Horné Krškany coded by families. Families at Nitra Horné Krškany do not cluster in dietary-specific groups. All plots are restricted to individuals with qpAdm estimates.

Figure 4:. Asparn-Schletz population size:

Inferred population size trajectory of Asparn-Schletz and Nitra based on HapNe-LD. The recent contraction in Nitra Horné Krškany likely reflects undetected families in the sample, while the Asparn-Schletz individuals have no evidence of being more closely related to each other than they are to the more widely sampled LBK. Error bars represent one (dark) and two (light) standard deviations.

Figure 5:. The LBK/ALPC networks:

I: A) A heatmap showing the intensity of IBD, presenting the average total length of IBD segments > 12cM shared between all possible pairs by area or period. The numbers after the site names show the number of individuals per site included in these analyses. B) Regression of summed IBD >12cM shared between individuals of each pair of sites (averaged over all pairs), and geographic distance. Polgár-Ferenci-hat has more connections with closer sites supporting a localised ALPC community, while Asparn-Schletz and Nitra Horné Krškany do not show a clear association with distance, as would be expected if the western LBK expansion was so rapid that nearby groups were hardly more closely related than groups far apart.

Figure 6:. Selection in Neolithic genomes:

(A) B1 scores in the ALPC. (B) B1 scores in the LBK. B1 shows regions with balancing selection, the highest signal on chromosome 6 at HLA.