The Red Fox Y-Chromosome in Comparative Context

Abstract

While the number of mammalian genome assemblies has proliferated, Y-chromosome assemblies have lagged behind. This discrepancy is caused by biological features of the Y-chromosome, such as its high repeat content, that present challenges to assembly with short-read, next-generation sequencing technologies. Partial Y-chromosome assemblies have been developed for the cat (Feliscatus), dog (Canislupusfamiliaris), and grey wolf (Canislupuslupus), providing the opportunity to examine the red fox (Vulpesvulpes) Y-chromosome in the context of closely related species. Here we present a data-driven approach to identifying Y-chromosome sequence among the scaffolds that comprise the short-read assembled red fox genome. First, scaffolds containing genes found on the Y-chromosomes of cats, dogs, and wolves were identified. Next, analysis of the resequenced genomes of 15 male and 15 female foxes revealed scaffolds containing male-specific k-mers and patterns of inter-sex copy number variation consistent with the heterogametic chromosome. Analyzing variation across these two metrics revealed 171 scaffolds containing 3.37 Mbp of putative Y-chromosome sequence. The gene content of these scaffolds is consistent overall with that of the Y-chromosome in other carnivore species, though the red fox Y-chromosome carries more copies of BCORY2 and UBE1Y than has been reported in related species and fewer copies of SRY than in other canids. The assignment of these scaffolds to the Y-chromosome serves to further characterize the content of the red fox draft genome while providing resources for future analyses of canid Y-chromosome evolution.

Keywords: BCORY2; MSY; UBE1Y; Vulpes vulpes; Y-chromosome; Y-chromosome genes; carnivore; copy-number variation; next-generation sequencing; sex chromosomes.

Figure A1

Cluster assignment of windows based on each metric individually. Cluster number is random, but clusters are color-coded to indicate whether they are most consistent with the X-chromosome (orange), autosomes (black), or Y-chromosome (blue).

Figure A2

Depth of coverage of scaffolds included in cluster 1 that were assigned to an autosome based on P_VSC_UK alone. Depth in male and female resequencing reads is indicated separately. Depth was averaged in males and females in windows of 10% of the scaffold’s length. Axes vary depending on the observed depth (y) and the length of the scaffold (x).

Figure 1

Average per-fox depth in males and females along dog chromosome X (CFAX). Bars represent the standard deviation of mean estimates, while the grey shading indicates the confidence interval based on smoothing with geom_smooth() in ggplot2 [51]. (A) Averages were calculated in 1-Mbp intervals. (B) Averages were calculated in 100-Kbp windows and are shown only for the region from chrX 6 Mbp to 7 Mbp. The two genes flanking the pseudoautosomal boundary on CFAX, SHROOM2 (pseudoautosomal) and WWC3 (X-chromosomal), are indicated as rectangles along the x-axis.

Figure 2

The distribution of male-specific sequence among scaffolds. This histogram indicates the percent of valid, single-copy k-mers unmatched in the female reads across all scaffolds evaluated.

Figure 3

Comparison of male and female coverage of windows. (A) In each window, the percent of the mapped reads coming from the male and from the female resequencing data was calculated. Due to the depth of coverage in the male resequencing data being slightly higher, there is some displacement of the autosomal and X-chromosome curves off 50% and 33/67%, respectively. The expected pattern of density distribution is apparent, with males and females contributing roughly equal numbers of reads to most windows. Very small peaks are apparent at 0% and 100% that correspond to primarily male contribution. (B) Zoomed-in depiction of only the small peaks that represent 75% or more contribution from a single sex.

Figure 4

Based on copy number variation (CNV) between males and females and male-specific sequence motifs, the windows form three clusters. The percent of reads mapping to each window that originated in males (as estimated with CNV-Seq) was plotted against the percent of male-specific 18-mers comprising each scaffold (as estimated with YGS.pl). Percentages were normalized based on standard score. (A) Windows from scaffolds with predicted chromosomal origins are plotted, and chromosomal origin is indicated by color. (B) Windows from all 12,625 scaffolds analyzed are included, and chromosomal origin, when known, is indicated by color. (C) Each window is color-coded according to the cluster to which it was assigned by k-means clustering.

Figure 5

Depth of coverage of Y-scaffolds by male and female resequencing reads. Depth was averaged in males and females over 5000-bp windows for scaffolds longer than 50 Kbp, or for windows 10% of the scaffold length for shorter scaffolds. Microsatellite marker positions [28] are indicated along the top with stars; gene positions as identified in Table 3 are indicated along the bottom. Only scaffolds containing genes are shown. Axes vary depending on the observed depth (y) and the length of the scaffold (x).

Figure 6

Estimated depth of coverage along BCORY1 on scaffold360 (A) and BCORY2 on scaffold310 (B). When the position of the codons could not be estimated from the BLAST results, the position is indicated below line at y = 0. Male and female depth is indicated separately. Each point represents an amino acid, with the depth estimated off of the surrounding 15 bp (approximately two codons on either side). Dotted lines indicate intervals of 2.5×, or approximately one copy.

Figure 7

Depth of coverage in the predicted exons of MSY genes. Based on research in dogs and cats [6], DDX3Y, USP9Y, UTY, and UBE1Y (top two rows) were predicted to be single-copy in males, whereas HSFY, SRY, and CUL4BY (bottom two rows) were predicted to be multi-copy in males. The depth of coverage suggests a single copy (as indicated by the lower dotted line) for all of these genes, except UBE1Y and CUL4BY, which appeared to be present at a high copy number. When a codon’s position could not be estimated from the BLAST results, that region of the amino acid is indicated below 0.