Reconstructing an African haploid genome from the 18th century

Datasets provided map the descendants in HJ's genealogy to identifiers used in the associated paper, African fragments in HJ descendants as determined by HAPMIX statistical software and the different African microarray SNP datasets used as reference populations to determine the origin of HJ’s haploid African genome.

Files are provided in either .xls (datasets) or .pdf (figures) format.

HAPMIX is an application for accurately inferring chromosomal segments of distinct ancestry in admixed populations, using dense genetic data. ADMIXTURE is a software tool for maximum likelihood estimation of individual ancestries from multilocus SNP genotype datasets.

A genome is a mosaic of chromosome fragments from different ancestors who existed at some arbitrary number of generations earlier. When chromosome fragments of one ancestor can be distinguished from those of others among a group of descendants, then his or her genome can be at least partially reconstructed. In the associated publication, we use this approach to reconstruct the genome of Hans Jonatan (HJ), born on the Caribbean island of St Croix in 1784 to an enslaved African mother and a European father. HJ migrated to Iceland in 1802, where he married an Icelandic woman and had two children. We genotyped 182 of his 788 descendants using single nucleotide polymorphism (SNP) chips and whole genome sequenced (WGS) 20 of them. Using these data, we reconstructed 38% of HJ’s maternal genome and inferred that his mother was most likely from the region spanned by Benin, Nigeria and Cameroon.

The associated analyses are undertaken to determine whether African fragments identified in descendants of HJ could be derived from other ancestors from the Icelandic population with African ancestry. We first ran HAPMIX for chip-typed 151,014 Icelanders (including the 182 descendants of HJ), using the phased CEU and YRI HapMap samples as source populations and parameters described in the methods. Consecutive African fragments were assembled as described in the methods. For this much larger set of putative African fragments, no further filtering was performed. This is because the filtering steps applied to the descendants of HJ were all predicated on the assumption of common descent from a single known ancestor with African ancestry and known genealogical paths. Such ancestors cannot be meaningfully postulated for the 150,832 genotyped Icelanders who are not descended from HJ. Consequently, we expect some small background level of false positive African fragments across all the genotyped Icelanders.