The reference genome of Miscanthus floridulus illuminates the evolution of Saccharinae

Abstract

Miscanthus, a member of the Saccharinae subtribe that includes sorghum and sugarcane, has been widely studied as a feedstock for cellulosic biofuel production. Here, we report the sequencing and assembly of the Miscanthus floridulus genome by the integration of PacBio sequencing and Hi-C mapping, resulting in a chromosome-scale, high-quality reference genome of the genus Miscanthus. Comparisons among Saccharinae genomes suggest that Sorghum split first from the common ancestor of Saccharum and Miscanthus, which subsequently diverged from each other, with two successive whole-genome duplication events occurring independently in the Saccharum genus and one whole-genome duplication occurring in the Miscanthus genus. Fusion of two chromosomes occurred during rediploidization in M. floridulus and no significant subgenome dominance was observed. A survey of cellulose synthases (CesA) in M. floridulus revealed quite high expression of most CesA genes in growing stems, which is in agreement with the high cellulose content of this species. Resequencing and comparisons of 75 Miscanthus accessions suggest that M. lutarioriparius is genetically close to M. sacchariflorus and that M. floridulus is more distantly related to other species and is more genetically diverse. This study provides a valuable genomic resource for molecular breeding and improvement of Miscanthus and Saccharinae crops.

Fig. 1. Genomic features of M. floridulus genome.

Tracks indicate the following: gene density (a); TE density (b); gene expression level (TPM; the tracks from outermost to innermost indicate root, stem, leaf and inflorescence respectively) (c); SNP/Indels density from the resequencing data of the 75 accessions (outer track: SNP density; inner circle: Indel density) (d). Chr, chromosome. The outer circle represents the chromosome length of M. floridulus, with units in Mb.

Fig. 2. Alignment of M. floridulus chromosomes with sorghum or S. spontaneum chromosomes.

a, A set of two homoeologous chromosomes aligned to a single sorghum chromosome except MfChr7, MfChr13 and MfChr8, among which MfChr7 aligned to Sbchr4, MfChr13 aligned to Sbchr7 and MfChr8 aligned to Sbchr4 and Sbchr7. b, Alignment of M. floridulus chromosomes with S. spontaneum chromosomes. c, The split chromosome homoeologous to Sbchr7 were inserted into the chromosome homoeologous to SbChr4 in M. floridulus. The red triangle indicates the centromere of MfChr8. Note that there is an inversion in Sbchr4 compared with MfChr7, MfChr8 (blue lines) and SsChr4ABCD. d, Gene count along the MfChr8 chromosome using 2 Mb as a window and 200 kb as a shift. Dashed lines indicate the joint regions linking the split chromosome segments homoeologous to Sbchr7 and Sbchr4 and the third region (blue dashed lines) corresponds to the new evolved centromere of MfChr8. Source data

Fig. 3. Evolutionary history of Saccharinae.

a, Distribution of synonymous nucleotide substitutions between inter- or intraspecies. The lines with different colours represents the Ks distribution of syntelogues between two species or subgenomes of same species. b, A schematic species tree outlining the evolutionary history of Saccharinae group. The dark green explosion shapes show WGD events and light blue triangles indicate genome rearrangement events. A time line (leftmost) is shown in Ma and a Ks line (rightmost) is also displayed. Chromosome numbers of each state of a plant are shown along branches. (1) Time of divergence of Miscanthus and Saccharum with Sorghum. (2) Time of divergence of Misanthus with Saccharum. (3) Time of divergence of Miscanthus subgenomes. (4) Time of autopolyploidization of the ancestor of S. spontaneum. The plot was based on our data and analyses plus part of the results from a previous paper. Source data

Fig. 4. The overview of cellulose synthases in M. floridulus.

a, Maximum likelihood phylogenetic tree of CesA genes from rice, sorghum, M. floridulus and S. spontaneum. Different colour ranges correspond to different CesA groups. Numbers at tree nodes represent bootstrap support values (1,000 replications). b, The syntenic relationship of CslH genes on Sbchr6, MfChr11, MfChr12 and Ss5ABCD. Grey lines indicate the homoeologous genes between any two syntenic regions. Note that there are 3, 5, 1, 3, 4, 2 and 2 homoeologues on the syntenic regions from Sbchr6, MfChr11, MfChr12, Ss5A, Ss5B, Ss5C and Ss5D, respectively. c, Heatmap of the expression level of CesA/Csl genes in different tissues (root, growing stem, mature leaf and inflorescence) of M. floridulus. The expression levels were shown as log₂(TPM + 0.5). Only genes with average TPM > 1 across all tissues were shown as indicated in Supplementary Data 3.

Fig. 5. Population genetic structure and phylogenetic relationships among 75 Miscanthus accessions.

a, Principal components (PCs) of accession variation. The percentage numbers in the brackets indicates the proportions of the two PCs. b, Bootstrapped tree of 75 Miscanthus accessions based on genetic distance. The tetraploid was labelled by the tail of the accession ID. c, ADMIXTURE plot for Miscanthus, showing the distribution of K = 3 genetic clusters with the smallest cross-validation error. Mg, M. × giganteus.