Analysis of muntjac deer genome and chromatin architecture reveals rapid karyotype evolution

Abstract

Closely related muntjac deer show striking karyotype differences. Here we describe chromosome-scale genome assemblies for Chinese and Indian muntjacs, Muntiacus reevesi (2n = 46) and Muntiacus muntjak vaginalis (2n = 6/7), and analyze their evolution and architecture. The genomes show extensive collinearity with each other and with other deer and cattle. We identified numerous fusion events unique to and shared by muntjacs relative to the cervid ancestor, confirming many cytogenetic observations with genome sequence. One of these M. muntjak fusions reversed an earlier fission in the cervid lineage. Comparative Hi-C analysis showed that the chromosome fusions on the M. muntjak lineage altered long-range, three-dimensional chromosome organization relative to M. reevesi in interphase nuclei including A/B compartment structure. This reshaping of multi-megabase contacts occurred without notable change in local chromatin compaction, even near fusion sites. A few genes involved in chromosome maintenance show evidence for rapid evolution, possibly associated with the dramatic changes in karyotype.

Fig. 1. Evolutionary and phylogenomic analyses.

a The phylogenetic tree of the five analyzed species, calculated from fourfold degenerate sites and divergence time confidence intervals, was visualized with FigTree (commit 901211e; https://github.com/rambaut/figtree). The ancestral karyotype at each node and the six branches with fission and fusion events relative to the ancestral karyotype were labeled on the tree. The lack of fissions or fusions on the R. tarandus-specific branch as well as the timings of the cervid-specific and B. taurus-specific fissions were derived from literature. b The alignment plot was generated with jcvi.graphics.karyotype (v0.8.12; https://github.com/tanghaibao/jcvi) using runs of collinearity containing at least 25 kb of aligned sequence between B. taurus, C. elaphus, M. reevesi, and M. muntjak. R. tarandus was excluded, as it is not a chromosome-scale assembly. Chromosomes that have been inverted in this image relative to their original assembly orientations are marked with asterisks. c Pairwise distances in substitutions per fourfold degenerate site extracted from the RAxML (v8.2.11) phylogenetic tree using Newick utilities (v1.6) were shown relative to the reference genome M. muntjak.

Fig. 2. Chromosome Hi-C contact maps.

Visualization of M. muntjak (below the diagonal) and M. reevesi (above the diagonal) Hi-C contact maps in Juicebox (v1.11.08) using the M. muntjak assembly as the reference. Orange boxes demarcate the boundaries of the three M. muntjak chromosomes, which are ordered as in Fig. 1. Chromosome numbers are provided in the lower-left corner of each. The intensity of blue pixels is proportional to the contact frequency between x and y pairs of genomic loci. The highest intensity pixels are along the diagonal of each chromosome, indicating a high degree of contacts between loci in close proximity. The checker board/striped patterns near the diagonal reflect fewer contacts between neighboring loci and increased contacts between more distant loci due to the three-dimensional chromatin folding (i.e., A/B compartment) structure within nuclei. In the upper triangle, the step-like pattern of high-density contacts along the diagonal is a result of conserved collinearity between M. reevesi and M. muntjak chromosomes; however, six blocks of high-frequency contacts (black arrows) can be observed off the diagonal and reflect large structural differences resulting from chromosome fission and fusion events. Two inverted segments (gray arrows) can also be observed.

Fig. 3. Evaluation of inter-chromosome contacts.

a Normalized 1 Mb intra-bin Hi-C contacts for M. muntjak (y-axis) vs. M. reevesi (x-axis) with the bins containing the M. muntjak lineage-specific fusion sites (Supplementary Table 7), chromosome ends, the X chromosome, the potential M. muntjak haplotype-specific duplication, and the potential M. muntjak haplotype-specific deletion colored. The expected result of conserved Hi-C contacts was represented with a dashed red line. For fusion site ranges spanning two bins, the bin containing the majority of the fusion site range was deemed to be the fusion site bin. b–d Copy number was calculated from normalized coverage of adapter-trimmed 10x Genomics linked reads for three regions with variation in the chromatin contacts: b the X chromosome, c the potential M. muntjak haplotype-specific duplication, and d the potential M. muntjak haplotype-specific deletion, with the copy number of M. muntjak in blue and M. reevesi in orange.