A high-resolution map of small-scale inversions in the gibbon genome

Abstract

Gibbons are the most speciose family of living apes, characterized by a diverse chromosome number and rapid rate of large-scale rearrangements. Here we performed single-cell template strand sequencing (Strand-seq), molecular cytogenetics, and deep in silico analysis of a southern white-cheeked gibbon genome, providing the first comprehensive map of 238 previously hidden small-scale inversions. We determined that more than half are gibbon specific, at least fivefold higher than shown for other primate lineage-specific inversions, with a significantly high number of small heterozygous inversions, suggesting that accelerated evolution of inversions may have played a role in the high sympatric diversity of gibbons. Although the precise mechanisms underlying these inversions are not yet understood, it is clear that segmental duplication-mediated NAHR only accounts for a small fraction of events. Several genomic features, including gene density and repeat (e.g., LINE-1) content, might render these regions more break-prone and susceptible to inversion formation. In the attempt to characterize interspecific variation between southern and northern white-cheeked gibbons, we identify several large assembly errors in the current GGSC Nleu3.0/nomLeu3 reference genome comprising more than 49 megabases of DNA. Finally, we provide a list of 182 candidate genes potentially involved in gibbon diversification and speciation.

Figure 1.

Strand-seq data analysis. (A) Schematic view of the method used to generate the Chr 13 composite file for the detection of the inversions. In the middle, an ideogram of human Chromosome 13 (HSA13) and the corresponding synteny blocks relative to Nomascus (NLE5 or NLE9) is shown. (Left) If single-cell Strand-seq data are pooled directly (Strand-seq output for six single cells is shown as an example) without taking into account the human–gibbon synteny block information, the resulting composite file will show no informative strand-state information (reads will map to both Watson and Crick strands equally, so every chromosome appears as WC). (Right) The method used in the present study considers human–gibbon synteny blocks individually to select libraries for generating a composite file separately for each syntenic region. Only informative libraries for each synteny block are selected and merged to generate a composite file of the whole chromosome. (B) Circos diagram (Krzywinski et al. 2009) reporting detected inversions for each human chromosome ideogram, with heterozygous inversions in the external circle and homozygous inversions in the inner circle. Green bars indicate the number of inversions detected for each synteny block (shown in different shades of green), where the height is proportional to the number of inversions detected for that block (min = 0; max = 9). (C) Circos diagram (Krzywinski et al. 2009) exemplifying the results of the analysis for a single human chromosome. Human–gibbon synteny blocks for human Chr 13 (large green highlights) and the inversions detected within each block (thin orange lines) are shown.

Figure 2.

Evolutionary history of inversions. The pie chart in the middle reports a summary of the evolutionary analyses for all the inversions detected by Strand-seq. The four histograms report the number of inversions that occurred in each lineage, with detail on what was already known from previous literature and what has been determined for the first time in the present study.

Figure 3.

High inversion rates in Nomascus revealed by distance-based and Bayesian analysis. (A) Evolutionary tree based on Markov chain Monte Carlo based on a set of 536 inversions considering both homozygous and heterozygous presence (see Methods). For each branch, the 95% posterior density confidence interval is reported. Numbers on nodes refer to shared inversion presence on the underlying subtrees, and numbers on branches refer to private inversions. (B) Upset plot of a set of 536 inversions. The upper barplot refers to the number of inversions observed in the species indicated in the intersection matrix. The left barplot refers to the total number of inversions identified for each species. Bar colors refer to the length of the inversions, as shown in the legend. Inversions putatively assigned to incomplete lineage sorting and/or recurrence were annotated with an asterisk. The silhouette of the chimpanzee is created by T. Michael Keesey and Tony Hisgett (PhyloPic; http://phylopic.org/; image is under a Creative Commons Attribution 3.0 unported license); silhouettes of bonobo and gorilla are from PhyloPic under a Public Domain Dedication 1.0 license, and the silhouette of the macaque is from PhyloPic under a Public Domain Mark 1.0 license.

Figure 4.

Gene density and repeat content at inversion breakpoints. (A) The panel shows data comparisons for LINE-1 between 10,000 random clusters of shuffled regions (gray area) and gibbon-specific BP regions devoid of SDs (dashed line) in human GRCh38. (B) The panel shows the gene content of 10,000 random clusters of shuffled regions, chosen in genomic segments with (green area) and without (orange area) SDs, compared with the gene content at gibbon-specific inversion BPs with (green dashed line) and without (orange dashed line) SDs.

Figure 5.

FISH validation of a 17-Mbp inversion. NSI Strand-seq view of NLE Chromosome 22a and HSA Chromosome 2 shows the incomplete switch in the orientation of 17 Mbp against both reference genomes, suggesting the presence of a heterozygous inversion (Chr2_inv16). Metaphase FISH experiments in human, NSI, and NLE show that the region is inverted in heterozygous state only in NSI and is in direct orientation in both human and NLE. (HSA) Homo sapiens; (NSI) Nomascus siki; (NLE) Nomascus leucogenys.