The dual origin of the Malagasy in Island Southeast Asia and East Africa: evidence from maternal and paternal lineages

Abstract

Linguistic and archaeological evidence about the origins of the Malagasy, the indigenous peoples of Madagascar, points to mixed African and Indonesian ancestry. By contrast, genetic evidence about the origins of the Malagasy has hitherto remained partial and imprecise. We defined 26 Y-chromosomal lineages by typing 44 Y-chromosomal polymorphisms in 362 males from four different ethnic groups from Madagascar and 10 potential ancestral populations in Island Southeast Asia and the Pacific. We also compared mitochondrial sequence diversity in the Malagasy with a manually curated database of 19,371 hypervariable segment I sequences, incorporating both published and unpublished data. We could attribute every maternal and paternal lineage found in the Malagasy to a likely geographic origin. Here, we demonstrate approximately equal African and Indonesian contributions to both paternal and maternal Malagasy lineages. The most likely origin of the Asia-derived paternal lineages found in the Malagasy is Borneo. This agrees strikingly with the linguistic evidence that the languages spoken around the Barito River in southern Borneo are the closest extant relatives of Malagasy languages. As a result of their equally balanced admixed ancestry, the Malagasy may represent an ideal population in which to identify loci underlying complex traits of both anthropological and medical interest.

Figure 1

A, Y-chromosomal haplogroup frequencies in the Malagasy and in potential ancestral populations. The maximum parsimony phylogeny relating the 41 Y-chromosomal haplogroups defined in the present study is shown above the absolute frequencies of those lineages in the different populations. The phylogeny is labeled with single-letter clades, and branches are labeled with the markers that define them. The lineage nomenclature is the updated version of that proposed by the YCC (YCC ; Jobling and Tyler-Smith 2003). The gray shading on the phylogeny indicates the nine sets of markers typed together in multiplexes. B, Pie charts illustrating the relative frequencies of the different haplogroups (colored to agree with the coloring of the phylogeny) shown on a map of the Indian Ocean. The Taiwanese population sample represents individuals pooled from four different aboriginal groups. Samples in these Island Southeast Asian and Pacific populations that were identified elsewhere as representing recent paternal European admixture in published lower-resolution Y-chromosomal marker typing of the same populations (Hurles et al. 2002) were not typed in the present study. East African data come from Luis et al. (2004).

Figure 2

A, Phylogeny of mt sequence types found in the Malagasy and their CoGs. The maximum parsimony phylogeny of the 14 maternal lineages defined in the present study builds on the phylogeny constructed by Kivisild et al. (2002). The tips of the phylogeny are labeled with the number of Malagasy mt genomes found in each lineage. The phylogeny is also labeled with the major clades and the variable sites that define individual branches. The gray shading on the phylogeny indicates the seven coding variants that are typed together in a single multiplex. B, A map of the Old World, showing the positions of the CoGs of the 14 different mt sequence types (blackened circles, triangles, and squares) and the lineage group to which that sequence type belongs. Strictly speaking, mtDNA type “L” encompasses mtDNA types “M” and “N,” which are Eurasian subgroups of “L”; however, for simplicity, “L” here denotes “L” lineages excluding “M” and “N” types. The number within each shape indicates the frequency of that lineage within the data set. For most mt types, it was sufficient to enter only their HVSI sequences. However, to obtain monophyletic hits in the geographic database for two Malagasy HVSI sequence types found in paraphyletic mt lineages, it was also necessary to consider their coding SNP haplotypes, to eliminate spurious matches in geographical regions in which these haplotypes are known to be absent. Full HVSI sequence types are given in table 3.