Unveiling the Genetic History of the Maniq, a Primary Hunter-Gatherer Society

Abstract

The Maniq of southern Thailand is one of the last remaining practicing hunter-gatherer communities in the world. However, our knowledge on their genetic origins and demographic history is still largely limited. We present here the genotype data covering ∼2.3 million single nucleotide polymorphisms of 11 unrelated Maniq individuals. Our analyses reveal the Maniq to be closely related to the Semang populations of Malaysia (Malay Negritos), who altogether carry an Andamanese-related ancestry linked to the ancient Hòabìnhian hunter-gatherers of Mainland Southeast Asia (MSEA). Moreover, the Maniq possess ∼35% East Asian-related ancestry, likely brought about by recent admixture with surrounding agriculturist communities in the region. In addition, the Maniq exhibit one of the highest levels of genetic differentiation found among living human populations, indicative of their small population size and historical practice of endogamy. Similar to other hunter-gatherer populations of MSEA, we also find the Maniq to possess low levels of Neanderthal ancestry and undetectable levels of Denisovan ancestry. Altogether, we reveal the Maniq to be a Semang group that experienced intense genetic drift and exhibits signs of ancient Hòabìnhian ancestry.

Keywords: Negritos; Semang; Southeast Asia; genetic ancestry; population genetics.

Fig. 1.

Maniq form a genetic cluster close to Malay Semang populations. (A) Map with approximate locations of all Asian populations included in the study. (B) Eigenvector 1 and 2 of a PCA on a worldwide set of populations, number of SNPs used = 320,032. (C) Eigenvector 1 and 2 of a PCA restricted to Asian populations, number of SNPs used = 319,845.

Fig. 2.

Admixture analysis for Asian populations. admixture plots for K = 2–7 with 170,513 SNPs used (K = 7 has the lowest cross-validation error).

Fig. 3.

High levels of genetic differentiation in the Maniq. (A) Total length of ROH versus total number of ROH per individual and (B) Number of ROH per segment length category covering the Yoruba, Dai, Maniq, Papuan, and Onge populations (for additional populations see supplementary figs. 4AandB, Supplementary Material online). Number of SNPs used = 779,956.

Fig. 4.

Different ancestry components of the Maniq. Results of f4 statistics including standard errors for (A) f4(Mbuti, Lao Hòabìnhian; Liangdao 2, X), (B) f4(Mbuti, Lao Hòabìnhian; Balangao, X), (C) f4(Mbuti, Liangdao 2; Lao Hòabìnhian, X), and (D) f4(Mbuti, Balangao; Onge, X), (E) f4(Mbuti, Papuan; Balangao, X) and (F) f4(Mbuti, Onge; Balangao, X) sorted from highest to lowest f4 values. A line marked with an asterisk indicates an absolute Z value of greater than 3. Number of SNPs used for (A–C) = 16,428, for (D–F) = 319,386. Please see supplementary table 4, Supplementary Material online, for the exact results including standard errors of (D).

Fig. 5.

Estimated admixture proportions of Andamanese-related and East Asian-related ancestries of the Semang Populations of MSEA. Pie chart results of qpAdm with 319,386 SNPs used. The position of the pie charts corresponds to the approximate location of their respective populations (Aghakhanian et al. 2015). Pie chart positions were slightly adjusted to prevent overlapping, notably the Mendriq were shifted to the right.

Fig. 6.

Genetic relationship of the Maniq relative to worldwide set of populations. TreeMix analysis with three added migration edges (over 99% explained variance). Number of SNPs used = 190,283.

Fig. 7.

Inferred admixture graph model for the Maniq. A possible qpGraph scenario to explain the topology around the Maniq. Worst fitting Z = 0.788, number of SNPs used = 56,590.