A principal obstacle to completing maps and analyses of the human genome involves the genomes inaccessible regions: sequences(often euchromatic and containing genes) that are isolated from the rest of the euchromatic genome by heterochromatin and otherrepeat-rich sequence. We describe a way to localize these sequences by using ancestry linkage disequilibrium in populations that derive ancestry from at least three continents, as is the case for Latinos. We used this approach to map the genomic locations of almost 20mega-bases of sequence unlocalized or missing from the current human genome reference (NCBIGenomeGRCh37)a substantial fraction of the human genomes remaining unmapped sequence. We show that the genomic locations of most sequences that originated from fosmids and larger clones can be admixture mapped in this way, by using publicly available whole-genome sequence data. Genomeassembly efforts and future builds of the human genome reference will be strongly informed by this localization of genes and othere uchromatic sequences that are embedded within highly repetitive pericentromeric regions.