Chromosome-scale genome assembly for the duckweed Spirodela intermedia, integrating cytogenetic maps, PacBio and Oxford Nanopore libraries

Abstract

Duckweeds are small, free-floating, morphologically highly reduced organisms belonging to the monocot order Alismatales. They display the most rapid growth among flowering plants, vary ~ 14-fold in genome size and comprise five genera. Spirodela is the phylogenetically oldest genus with only two mainly asexually propagating species: S. polyrhiza (2n = 40; 160 Mbp/1C) and S. intermedia (2n = 36; 160 Mbp/1C). This study combined comparative cytogenetics and de novo genome assembly based on PacBio, Illumina and Oxford Nanopore (ON) reads to obtain the first genome reference for S. intermedia and to compare its genomic features with those of the sister species S. polyrhiza. Both species' genomes revealed little more than 20,000 putative protein-coding genes, very low rDNA copy numbers and a low amount of repetitive sequences, mainly Ty3/gypsy retroelements. The detection of a few new small chromosome rearrangements between both Spirodela species refined the karyotype and the chromosomal sequence assignment for S. intermedia.

Figure 1

Six linkages due to chromosome rearrangements between S. polyrhiza and S. intermedia (clone 8410) are also present in the sequenced clone 7747. (a) ChrSp05–ChrSp06 = ChrSi06; (b) ChrSp08–ChrSp18 = ChrSi09; (c) ChrSp03–ChrSp17 = ChrSi04; (d) ChrSp10–ChrSp16 = ChrSi11; (e) ChrSp06–ChrSp07–ChrSp14 = ChrSi07; (f) ChrSp03–ChrSp06–ChrSp14 = ChrSi03. See also Fig. 3. Scale bars = 5 µm.

Figure 2

Rearrangements between S. polyrhiza (n = 20) and S. intermedia (n = 18), confirmed for chromosomes of clones 8410 and 7747. Enframed: newly found rearrangement. Red boxes: present only in S. polyrhiza; green boxes: sequences present only in S. intermedia. Scale bar = 3 Mbp (based on PacBio assembly for clone 7747). Enumeration is as in Hoang & Schubert (2017) and in Table S3 (Short sequences present in S. p., but not chromosomally assigned in S. i. correspond to “SiUn” and are not considered). 13 ChrSi of the 7747 assembly show telomeric sequences at one (*) or both ends (**), while all pseudomolecules of the 8410 assembly show them at both ends.

Figure 3

Circos plot of genomes of S. polyrhiza 9505 (orange) and S. intermedia 8410 (blue). (A) tracks representing the size of the pseudomolecules with a corresponding scale in 1 Mbp steps, with highlights every 5 Mbps, (B) total length of repeat features (in kbps) (C) gene density and (D) pairwise sequence synteny. The synteny link between gi|13 and Si02 is based on an error in the ON assembly for clone 9509 (see Fig. 5). The region in question actually belongs to gi|2 as does the remaining part of Si02. Gene and repeat density are plotted in 0.5 Mbps bins. Data for S. polyrhiza 9505 are from Michael et al. 2017.

Figure 4

New rearrangement between ChrSp 20 and ChrSp 15 in S. intermedia (a) The newly tested BAC 013O04 (yellow) belongs to ChrSp 20 (green) in S. polyrhiza; (b,c) BAC 013O04 was translocated to ChrSp 15 (red) forming ChrSi 16 in S. intermedia clones 8410 and 7747. Scale bars = 5 µm.

Figure 5

Evidence for ON mis-assembly of chromosome 13 of S. polyrhiza 9509. The newly tested BACs 029A10, 028I16 and 037J09 (yellow) belong to ChrSp 02 (red) of S. polyrhiza clone 9509 (upper panel) and clone 7498 (lower panel), not to ChrSp13. Scale bars = 5 µm.

Figure 6

Schematic representation of rDNA loci in S. intermedia 8410. (a) Depiction of 45S rDNA loci at chromosomes ChrSi01 and ChrSi06; Tel = telomere. (b) Depiction of 5S rDNA loci at chromosome ChrSi15, and ChrSi14. The ChrSi15 locus contains 31 5S rDNA units with NTS of type-1; the ChrSi14 locus is composed of two clusters of 7 and 13 5S rDNA units of type-2 NTS, separated by a doubled 6 kb sequence of unknown function.