High-throughput single-molecule mapping links subtelomeric variants and long-range haplotypes with specific telomeres

Abstract

Accurate maps and DNA sequences for human subtelomere regions, along with detailed knowledge of subtelomere variation and long-range telomere-terminal haplotypes in individuals, are critical for understanding telomere function and its roles in human biology. Here, we use a highly automated whole genome mapping technology in nano-channel arrays to analyze large terminal human chromosome segments extending from chromosome-specific subtelomere sequences through subtelomeric repeat regions to terminal (TTAGGG)n repeat tracts. We establish detailed maps for subtelomere gap regions in the human reference sequence, detect many new large subtelomeric variants and demonstrate the feasibility of long-range haplotyping through segmentally duplicated subtelomere regions. These features make the method a uniquely valuable new tool for improving the quality of genome assemblies in complex DNA regions. Based on single molecule mapping of telomere-terminal DNA fragments, we provide proof of principle for a novel method to estimate telomere lengths linked to distinguishable telomeric haplotypes; this single-telomere genotyping method may ultimately enable delineation of human cis elements involved in telomere length regulation.

Figure 1.

Consensus genome maps of individual subtelomeric regions containing SRE regions of NA12878. (A) Human subtelomere reference assemblies as described in Stong et al. (23) for 19p, 15q and 6p. Subtelomeric Repeat Element (SRE) paralogy blocks are shown in colors, with the gray line segments indicating adjacent 1-copy subtelomere DNA. The 19p and 15q assemblies extend to the start of the telomere terminal repeat (TTAGGG)n tract on the left (T), whereas there is a gap of indeterminate size between the start of the 6p reference sequence and the telomere. (B) Consensus single-molecule maps from the NA12878 genome aligning to the 19p, 15q and 6p telomeres of the Stong et al. (23) reference assemblies. In the consensus map for each subtelomere, nicking sites are represented by dark blue vertical lines within the light blue rectangles above the single DNA molecule maps. Each yellow row represents a single mapped large genomic DNA molecule, with the green ticks on each yellow row showing the labeled nicking sites imaged on that molecule. Lighter green ticks indicate nicking sites that did not match to the reference sequence. The numbers above the 19p consensus map indicate distance in megabases (Mb). The red bars in Panel B show SRE paralogy region blocks 1–5 from panel A that are shared by 19p and 15q. The dark blue bars in Panel B show SRE paralogy region blocks 4–8 from Panel A that are shared between 19p and 6p as well as part of 15q. The green bars in Panel B show a DNA segment shared between 19p and 6p but not 15q. The purple bar in Panel B shows a structurally variant segment of the 15q subtelomere, an insertion of about 50 kb immediately adjacent to the telomere relative to the reference sequence. The black bar in Panel B indicates the 6p gap region delineated by the consensus map.

Figure 2.

Haplotype-resolved subtelomeric variants of 15q. SRE paralogy blocks from Stong et al. reference assembly (23) are shown in colors, with the gray line segments indicating adjacent 1-copy subtelomere DNA. The 15q assembly begins at the centromeric end of the telomere terminal repeat (TTAGGG)n tract on the left (T). The hg38 representation of this assembly, which is identical to that of Stong et al. (23) except for the addition of a 10 kb gap adjacent to the telomere to represent unsequenced (TTAGGG)n (see supplementary materials), has been in silico nicked with the Nicking enzyme Nt.BspQI used for genome-wide mapping (represented by dark blue vertical lines within the light blue rectangles). Consensus single-molecule maps from the NA12892, 12891 and 12878 genomes aligning to the 15q telomeres are represented as described in Figure 1, and positioned beneath the hg38 reference map. Each yellow row represents a single molecule, with nicking sites shown in green. NA12892 contains haplotypes 1 and 2, while NA12891 has a third haplotype and NA12878 has haplotypes 1 and 3. The red arrows indicate the position of a nicking site that is present in the third haplotype but absent in the first and second haplotypes.

Figure 3.

Gap characterization and subtelomeric haplotype structures of 6p. SRE paralogy blocks from Stong et al. (23) are shown in colors, with the gray line segments indicating adjacent 1-copy subtelomere DNA. The 6p subtelomere assembly contains a gap of indeterminate size between the telomere and the start of the assembly. This gap is omitted in the Stong et al. (23) reference, and represented in the hg38 reference by a telomere (TTAGGG)n gap (10 kb) plus a clone gap (50 kb) on the telomeric side of the 6p assembly start position (see Supplementary Data); the hg38 reference is otherwise identical to the Stong et al. (23) reference sequence. In silico nicking of the hg38 reference with the Nicking enzyme Nt.BspQI used for genome-wide mapping is shown (represented by dark blue vertical lines within the light blue rectangles); note that the subtelomere gap region of hg38, a 60 kb string of NNN's in the sequence itself, has no nicking sites represented in the electronic digest. Consensus single-molecule maps from the NA12892, 12891 and 12878 genomes aligning to the 6p telomere are represented as described in Figure 1, and positioned beneath the hg38 reference map. The position of the gap region of the hg38 reference sequence is represented above each consensus map by a black rectangle. NA12892 shows two structurally variant haplotypes, as does NA12891, but NA12878 has inherited only haplotype 1 from each parent. While structurally variant within the gap region of 6p, the two haplotypes clearly diverge ∼120 kb from the telomere within SRE paralogy region 5.

Figure 4.

Detection of internal subtelomeric structural variations. The in-silico nicked 2p subtelomere region of the hg38 assembly is shown directly above consensus single-molecule maps of this region (represented by the yellow bar with green ticks) for each of the HG02490, HG02606 and HG02522 genomes. Note the 10 kb gap region represented at the start of hg38 (see Supplementary Data). Both HG02490 and HG2606 show an insertion at coordinate 0.2Mb. The blue bar here is the reference, hg38 and the yellow bar represents the consensus map of the molecules. Light gray lines connect the green dashes, which indicate nicking sites, to their locations on hg38. At 0.2Mb an insertion occurs on both HG02606 and HG02490.

Figure 5.

Telomere length measurement. The strategy of telomere length estimation through single-molecule mapping is shown in (A and B). Panel A shows a single DNA molecule (yellow line) that is aligned to the reference sequences (blue line). The telomere end (dotted red line) of the DNA molecule entered the nano-channel first. The three fluorescent labels (green dots on the yellow line) on the DNA molecule are mapped to the sequence motifs on the reference (green dots on the blue line). Panel B shows the same molecule that entered the nano-channel in a different orientation, with the telomere end entering the last. Panel C shows the single molecules used to estimate the telomere lengths from chromosome 5p of NA12878 and NA12891.