Y-Chromosomal DNA Variation in Pakistan

TL;DR

Y-chromosome binary and STR markers in 718 Pakistani males from 12 ethnic groups identify 11 haplogroups and 503 haplotypes, with frequencies broadly similar to neighboring regions but substantial population substructure. The Y data support Iranian origin of the Parsis, Mongolian/Genghis Khan–related descent of the Hazaras, and an African contribution to the Negroid Makrani, but not Tibetan, Syrian, Greek, or Jewish origins for other tested populations.

Key findings

• Among 718 males, 18 binary polymorphisms and 16 Y-STRs identified 11 stable haplogroups and 503 combination binary-marker/STR haplotypes; 5 haplogroups accounted for 92% of the sample.
• Haplogroup frequencies were generally similar to neighboring areas, and the language-isolate Burusho, Dravidian-speaking Brahui, and Sino-Tibetan-speaking Balti resembled the Indo-European-speaking majority; median-joining networks nevertheless showed population-specific clusters and substructuring.
• Admixture estimates supported external contributions for Hazaras from Mongolia (Long’s WLS .67, mR .52, mρ .41), Kalash from Greece (.4, .32, .23), Negroid Makrani from sub-Saharan Africa (.12, .12, .13), and Parsis from Iran (1.21, 1.06, 1), but not for Baluch from Syria, Balti from Tibet, Burusho/Pathan from Greece, or Kashmiri/Pathan from Jews (estimates 0 or negative).
• BATWING TMRCA estimates included Pakistani haplogroup 28 at ~7,000 years (95% CI 4,000–14,000), the Parsi-specific haplogroup 28 cluster at 1,800 years (600–4,500), and Hazara-specific haplogroup 1 and 10 clusters at 400 years (120–1,200) and 100 years (6–600), respectively.

Methods

• The study analyzed 718 unrelated male subjects from 12 ethnic groups of Pakistan; DNA was extracted from Epstein-Barr virus–transformed lymphoblastoid cell lines, with informed consent obtained.
• Samples were typed for 18 binary Y-chromosome markers (15 SNPs, an Alu insertion, a LINE1 insertion, and the 12f2 deletion) and 16 multiallelic Y-STR loci.
• Analyses included principal-components analysis, admixture estimation by Long’s weighted least-squares, mR, and mρ, AMOVA/population pairwise F_ST, median-joining networks, and BATWING TMRCA estimation.

Limitations

• The paper states that some population samples were small (e.g., Balti n=13, Kashmiri n=12, Hazara n=23), that origin conclusions assume modern source populations are representative of ancient populations, and that BATWING confidence intervals account for mutation-rate, growth, and subdivision uncertainty but not generation-time uncertainty.
• Evident from the text: the Punjabi-speaking majority was not analyzed, only 12 ethnic groups were included, some haplogroups were rare or absent (with F_ST set to 1 in some pairwise comparisons), and admixture estimates could be negative or zero, indicating limited resolution.
• The study uses only Y-chromosome data, so it captures male-specific history and cannot by itself resolve maternal or autosomal contributions; the authors note that mitochondrial DNA and other loci would help clarify histories such as those of the Negroid Makrani and Balti.