Novel insights on demographic history of tribal and caste groups from West Maharashtra (India) using genome-wide data

Abstract

The South Asian subcontinent is characterized by a complex history of human migrations and population interactions. In this study, we used genome-wide data to provide novel insights on the demographic history and population relationships of six Indo-European populations from the Indian State of West Maharashtra. The samples correspond to two castes (Deshastha Brahmins and Kunbi Marathas) and four tribal groups (Kokana, Warli, Bhil and Pawara). We show that tribal groups have had much smaller effective population sizes than castes, and that genetic drift has had a higher impact in tribal populations. We also show clear affinities between the Bhil and Pawara tribes, and to a lesser extent, between the Warli and Kokana tribes. Our comparisons with available modern and ancient DNA datasets from South Asia indicate that the Brahmin caste has higher Ancient Iranian and Steppe pastoralist contributions than the Kunbi Marathas caste. Additionally, in contrast to the two castes, tribal groups have very high Ancient Ancestral South Indian (AASI) contributions. Indo-European tribal groups tend to have higher Steppe contributions than Dravidian tribal groups, providing further support for the hypothesis that Steppe pastoralists were the source of Indo-European languages in South Asia, as well as Europe.

Figure 1

ROH analysis of the six West Maharashtrian populations. Deshastha Brahmins caste (DB) in red, Kunbi Marathas caste (KM) in violet, Kokana tribe (KN) in blue, Pawara tribe (PW) in yellow, Bhil tribe (BH) in green, and Warli tribe (WR) in orange. As admixed populations with no inbreeding we use ASW (Americans of African Ancestry in SW, USA) and ACB (African Caribbeans in Barbados) in turquoise from the 1KG cohort. (A) Violin plots showing the distribution of ROH within populations for the median total sum of ROH, median total sum of ROH shorter than 1.5 Mb and median total length of ROH longer than 1.5 Mb (white dots represent medians). (B) Mean total length of ROH (Mb) over six classes of ROH tract lengths. ROH classes: 0.3 < ROH < 0.5 Mb, 0.5 < ROH < 1 Mb, 1 < ROH < 2 Mb, 2<ROH < 4 Mb, 4<ROH < 8 Mb and ROH > 8 Mb. (C) Mean number of ROH and sum of ROH is plotted for each individual included in this study. The perpendicular broken red lines in at X = 25, X = 115 and X = 267 represent thresholds for second cousin, first cousin and uncle and niece offspring inbreeding coefficients. (D) Population Analysis and components of the inbreeding coefficient. The median of the systematic inbreeding (F_IS) versus the median of the genomic inbreeding coefficient (F_ROH) is shown for each population. Diagonal broken line represents F_IS = F_ROH. Horizontal broken line represents F_IS = 0. IQR (Interquartile range) is shown for both F_IS and F_ROH in each population.

Figure 2

PCA plot including the tribal and caste groups from West Maharashtra, as well as relevant modern and ancient DNA samples. The ancient DNA samples were projected onto the modern DNA samples.

Figure 3

Unsupervised clustering analysis using the program ADMIXTURE using K = 3, including tribal and caste groups from West Maharashtra, as well as relevant modern and ancient DNA samples.