Post by Admin on Jul 3, 2022 17:40:40 GMT
Evidence for at least five streams of migration into Micronesia
To determine the minimum number of migration streams into Micronesia needed to explain the data, we computed a statistic f4(X, New Guinea Highlanders; Kankanaey Igorot, Australian) proportional to FRO ancestry and correlated it to statistics sensitive to different types of East Asian and Papuan-associated ancestry (9). We identified at least five distinct migratory streams, as follows.
(M1 to M3) Three streams of FRO migration into Micronesia including a previously unknown lineage. We plotted a statistic measuring affinity to the two previously identified (7, 28) lineages FROSouthwestPacific and FROMarianas, specifically, f4(X, New Guinea Highlanders; Lapita, Unai) against our statistic measuring overall FRO ancestry proportion. All populations from the southwest Pacific and Polynesia fall on a line with a positive slope, implying closer affinity to Lapita than to Unai consistent with the Lapita-associated lineage being the source of their East Asian–associated ancestry (all residuals |Z| < 2 after regression; Fig. 3B and fig. S9B). Individuals from Central Micronesia (Pohnpei and Chuuk, and some other present-day Micronesians) also closely track the line (all residuals |Z| < 2), suggesting FRO ancestry from the Lapita expansion. By contrast, present-day individuals from Palau and the Mariana Islands yield negative f4-statistics (all residuals |Z| > 4), implying FRO sources less closely related to the Lapita individuals (tables S12 and S13). We confirmed with f4-symmetry statistics that all the prehistoric Remote Oceanian groups with nearly entirely East Asian–associated ancestry (Lapita, Unai, and Latte) descended from a common ancestral FRO population (table S22), which split earlier from the ancestors of indigenous and Iron Age Taiwanese and even earlier from those of Kankanaey Igorot. A surprise is that despite the fact that the Latte and Unai individuals share more alleles with each other than either group does with Lapita, there is not a simple tree relating these three groups, with the statistic f4(Latte, Unai; Lapita, diverse East Asians) yielding many significant negative Z values (maximum |Z| > 4; table S26). This suggests that the Latte individuals harbor admixture from a basal FRO lineage, which split from the lineages ancestral to Unai and Lapita before they separated from each other, a scenario that fits the data in explicit admixture graph modeling (Fig. 4C and figs. S12 to S15). We call this third lineage FROPalau because the proportion of this lineage is maximized in modern Palauans (where we estimate that it contributes 62% ancestry versus 15% in Latte individuals) (fig. S13A).
Fig. 3. Different Papuan and East Asian affinities.
(A and B) Test for differential (A) Papuan and (B) FRO affinities using a merge of the 1240K and MEGA data (~169,000 SNPs). Equation 1 (Eq. 1) is computed with all groups from Vanuatu and Polynesians, Eq. 2 with all Micronesian and New Guinea–related groups except those from Guam and Saipan, and Eq. 3 with all present-day groups except Micronesians. We show one standard error in each direction on the y axis. We merged Lapita individuals from Vanuatu and Tonga. See fig. S9 for the same analysis performed on individuals for whom we have ~397,000 SNPs genotyped on a merge of 1240K and Human Origins data.
To determine the minimum number of migration streams into Micronesia needed to explain the data, we computed a statistic f4(X, New Guinea Highlanders; Kankanaey Igorot, Australian) proportional to FRO ancestry and correlated it to statistics sensitive to different types of East Asian and Papuan-associated ancestry (9). We identified at least five distinct migratory streams, as follows.
(M1 to M3) Three streams of FRO migration into Micronesia including a previously unknown lineage. We plotted a statistic measuring affinity to the two previously identified (7, 28) lineages FROSouthwestPacific and FROMarianas, specifically, f4(X, New Guinea Highlanders; Lapita, Unai) against our statistic measuring overall FRO ancestry proportion. All populations from the southwest Pacific and Polynesia fall on a line with a positive slope, implying closer affinity to Lapita than to Unai consistent with the Lapita-associated lineage being the source of their East Asian–associated ancestry (all residuals |Z| < 2 after regression; Fig. 3B and fig. S9B). Individuals from Central Micronesia (Pohnpei and Chuuk, and some other present-day Micronesians) also closely track the line (all residuals |Z| < 2), suggesting FRO ancestry from the Lapita expansion. By contrast, present-day individuals from Palau and the Mariana Islands yield negative f4-statistics (all residuals |Z| > 4), implying FRO sources less closely related to the Lapita individuals (tables S12 and S13). We confirmed with f4-symmetry statistics that all the prehistoric Remote Oceanian groups with nearly entirely East Asian–associated ancestry (Lapita, Unai, and Latte) descended from a common ancestral FRO population (table S22), which split earlier from the ancestors of indigenous and Iron Age Taiwanese and even earlier from those of Kankanaey Igorot. A surprise is that despite the fact that the Latte and Unai individuals share more alleles with each other than either group does with Lapita, there is not a simple tree relating these three groups, with the statistic f4(Latte, Unai; Lapita, diverse East Asians) yielding many significant negative Z values (maximum |Z| > 4; table S26). This suggests that the Latte individuals harbor admixture from a basal FRO lineage, which split from the lineages ancestral to Unai and Lapita before they separated from each other, a scenario that fits the data in explicit admixture graph modeling (Fig. 4C and figs. S12 to S15). We call this third lineage FROPalau because the proportion of this lineage is maximized in modern Palauans (where we estimate that it contributes 62% ancestry versus 15% in Latte individuals) (fig. S13A).
Fig. 3. Different Papuan and East Asian affinities.
(A and B) Test for differential (A) Papuan and (B) FRO affinities using a merge of the 1240K and MEGA data (~169,000 SNPs). Equation 1 (Eq. 1) is computed with all groups from Vanuatu and Polynesians, Eq. 2 with all Micronesian and New Guinea–related groups except those from Guam and Saipan, and Eq. 3 with all present-day groups except Micronesians. We show one standard error in each direction on the y axis. We merged Lapita individuals from Vanuatu and Tonga. See fig. S9 for the same analysis performed on individuals for whom we have ~397,000 SNPs genotyped on a merge of 1240K and Human Origins data.