Phased chromosome-scale genome assembly of an asexual, allopolyploid root-knot nematode reveals complex subgenomic structure

Abstract

We present the chromosome-scale genome assembly of the allopolyploid root-knot nematode Meloidogyne javanica. We show that the M. javanica genome is predominantly allotetraploid, comprising two subgenomes, A and B, that most likely originated from hybridisation of two ancestral parental species. The assembly was annotated using full-length non-chimeric transcripts, comparison to reference databases, and ab initio prediction techniques, and the subgenomes were phased using ancestral k-mer spectral analysis. Subgenome B appears to show fission of chromosomal contigs, and while there is substantial synteny between subgenomes, we also identified regions lacking synteny that may have diverged in the ancestral genomes prior to or following hybridisation. This annotated and phased genome assembly forms a significant resource for understanding the origins and genetics of these globally important plant pathogens.

Fig 1. Genome profiling plots.

(A) Smudgeplot (left) proposing M. javanica as tetraploid, reporting the predicted percentages of ploidy levels in the genome as follows: tetraploid (48%), triploid (22%), or diploid (29%). (B) GenomeScope2 plot (right) showing four distinct peaks in both the predicted model of tetraploidy (black line) and in the observed k-mer spectra (blue fill). The amount of unique sequence falls to zero shortly after the fourth peak, indicating that k-mers at higher ploidies than four were mostly repetitive elements.

Fig 2. Ideogram and coverage depth of longest 33 scaffolds.

Scaffolds are coloured according to phasing status; blue—subgenome A, red—subgenome B, purple—unphased. A, Ideogram of 33 largest scaffolds. These 33 scaffolds contain 98% the total length of the assembly, with the remaining 36 contigs being shorter than 250 kbp, containing few gene models (0.19%), and consisting of mostly repetitive elements. B, Boxplot displaying the distribution of coverage depth for each scaffold. Red points denote the mean of data in each box. Coverage has been limited to a maximum of 800x to exclude probable repetitive sites with anomalous coverage depth. Dashed line shows the overall mean for all coverage levels across 33 scaffolds. Dotted line shows the mode of all coverage across 33 scaffolds.

Fig 3. Coverage depth frequency distributions of scaffolds 1, 2, 3, and 19.

X-axis represents coverage depth and y-axis represents frequencies of coverage. Colour indicates phase status: red for subgenome B, blue for subgenome A, and purple for unphased scaffolds. Bin size = 5. (A) Scaffold 1, top left, displays two main peaks of coverage with a small tail suggesting short collapsed regions. (B) Scaffold 2, top right, displays four peaks of coverage, indicating that much of this scaffold is collapsed and four copies are mapping to it. (C) Scaffold 3, bottom left, shows two peaks and very little tail, indicating two copies mapping and little to no assembly collapse. (D) Scaffold 19, bottom right, shows only one peak at ~120x coverage, suggesting that only one copy maps to this scaffold.

Fig 4. Macrosynteny analysis between subgenomes.

Glyphs along top and bottom represent scaffolds assigned to either subgenome A (blue) or subgenome B (red). Green lines mark locations of synteny between transcribed genes identified from mapping of Iso-Seq sequences. Grey lines mark locations of synteny between genes identified through MAKER3 gene prediction. Synteny and collinearity were identified using the MCScan module of JCVI using Iso-Seq informed transcriptional annotation. Scaffolds that could not be assigned to a subgenome are not shown.