Complete chromosome 21 centromere sequencing of families with Down syndrome

Down syndrome, the most common form of human intellectual disability, results from nondisjunction and an extra copy of chromosome 21 (chr21), also known as trisomy 21 (T21). Small centromeres have been hypothesized to contribute to its etiology, and studies on mice suggest that larger centromeres are more efficiently transmitted, yet complete sequencing of chr21 centromeres has been particularly challenging due to their repetitive nature and homology to chromosome 13. Using long-read sequencing, we sequenced and assembled the centromeres from eight families with a child with T21 (one parent-child trio, six mother-child duos, and one singleton), all resulting from maternal meiosis I errors. A comparison of all proband chr21 centromeres (n = 24) to those of control individuals (n = 287) shows that small centromeres are not enriched in families with T21 (p value = 0.72), contrary to earlier reports. However, chr21 extreme centromere size asymmetry (>10-fold) was observed for two of them. Mothers from these two families with T21 carry some of the smallest chr21 centromeres (143 and 181 kbp) observed in female individuals to date, exhibiting a ∼10.7- and ∼19.4-fold centromeric α-satellite higher-order repeat array size difference between the maternally inherited homologs, respectively. Phylogenetic reconstruction reveals that human chr21 is particularly prone to such asymmetry, with some of the biggest size differences occurring over the last ∼17,000 years of human evolution.