Alopecurus aequalis plants and materials

Seeds of Alopecurus aequalis (orange foxtail), donated by a private Individual in 2014 to the Royal Botanic Gardens, Kew Millennium Seed Bank (Serial Number 828127), were used for genome sequencing and annotation. While the collection location was not recorded, A. aequalis is not an agricultural weed in the UK, making it unlikely that it was collected from an active agricultural field. Consequently, these seeds were considered a herbicide-naive A. aequalis genotype. To establish a seed stock, 26 plants were grown and allowed to intrabreed in isolation for one generation at Rothamsted Research. For genome size estimation by flow cytometry, fresh leaf material from four individual plants (ID 828127A, B, C and D) of this second generation were analysed using the fluorochrome propidium iodide with the ‘one-step method’²³. Nuclei were isolated in the LB01 nuclei isolation buffer²⁴ and Petroselinum crispum ‘Champion Moss Curled’ was used as the internal calibration standard with an assumed 2C-value of 4.5 pg²⁵. The mean relative fluorescence of nuclei from A. aequalis and P. crispum were used to estimate the genome size of A. aequalis using the following equation: 2C-value of A. aequalis (pg) = (Mean peak position of A. aqualis / mean peak position of P. crispum) ⨯ 4.5.

To generate sufficient material for genome sequencing, a single plant (ID 828217A) was grown and vegetatively cloned twice following protocols described in the supplementary material in Cai et al. (2023)¹⁸. DNA for genome sequencing was isolated using the protocols described below from young leaves and meristem material from plants that had been dark-adapted for five days, flash frozen in liquid nitrogen, and stored at -80°C until shipping on dry ice. Similarly, RNA for Iso-Seq and RNA sequencing for annotation was sourced from flag leaves and flowering heads from clones of plants 828217A and 828217B. Additional RNA came from bulked shoot or root material derived from five technical replicates of 5-8 plants from the 828217 seed stock. These plants were grown under sterile hydroponic conditions, as per supplementary material of Cai et al. (2023), for a total of 42 days before separating root or shoot material from the seed. The flash frozen material was stored at -80°C until it was shipped on dry ice for processing. One clone from 828217A was allowed to flower and preserved by preparing a herbarium voucher which is stored at the Royal Botanic Garden Edinburgh (E) (Figure 1G, https://data.rbge.org.uk/herb/E01358418)

DNA extraction

HMW DNA extraction was performed using the Nucleon PhytoPure kit, with a slightly modified version of the manufacturer’s protocol. One gram of snap-frozen leaf material was ground under liquid nitrogen for 10 minutes. The tissue powder was thoroughly resuspended in Reagent 1 using a 10mm bacterial spreader loop until the mixture appeared completely homogeneous, at which point 4µl of 100mg/ml RNase A (Qiagen cat no. 19101) was added and mixed again. After incubation on ice, 200µl of resin was added along with chloroform. The chloroform extraction was followed by a phenol:chloroform extraction. Here, an equal volume of 25:24:1 phenol:chloroform:isoamyl alcohol was added to the previous upper phase, mixed gently at 4°C for 10 minutes and then centrifuged at 3000g for 10 minutes. The upper phase from this procedure was then transferred to another 15ml Falcon tube and precipitation proceeded as recommended by the manufacturer’s protocol. The final elution was left open in a chemical safety cabinet for two hours to allow residual phenol and ethanol to evaporate, and the DNA sample was left at room temperature overnight before further processing.

PacBio HiFi library preparation and sequencing

In total, 65µg of genomic DNA was split into 5 aliquots and manually sheared with the Megaruptor 3 instrument (Diagenode, P/N B06010003), according to the Megaruptor 3 operations manual, loading each aliquot of 150µl at 21ng/µl, with a shear speed of 31. Each sheared aliquot underwent AMPure® PB bead (PacBio®, P/N 100-265-900) purification and concentration before undergoing library preparation using the SMRTbell® Express Template Prep Kit 2.0 (PacBio®, P/N 100-983-900). The HiFi libraries were prepared according to the HiFi protocol version 03 (PacBio®, P/N 101-853-100) and the final libraries were size fractionated using the SageELF® system (Sage Science®, P/N ELF0001), 0.75% cassette (Sage Science®, P/N ELD7510), running on a 4.55hr program with 30µl of elution buffer per well post elution. The libraries were quantified by fluorescence (Invitrogen Qubit™ 3.0, P/N Q33216) and the size of the library fractions were estimated from a smear analysis performed on the FEMTO Pulse® System (Agilent, P/N M5330AA).

The loading calculations for sequencing were completed using the PacBio® SMRT®Link Binding Calculator 10.2. Sequencing primer v2 was annealed to the adapter sequence of the HiFi libraries. The libraries were bound to the sequencing polymerase with the Sequel® II Binding Kit v2.0. Calculations for primer and polymerase binding ratios were kept at default values for the library type. Sequel® II DNA internal control 1.0 was spiked into each library at the standard concentration prior to sequencing. The sequencing chemistry used was Sequel® II Sequencing Plate 2.0 (PacBio®, P/N 101-820-200) and the Instrument Control Software v 10.1. The libraries were sequenced on the Sequel IIe across 15 Sequel II SMRT®cells 8M. The parameters for sequencing per SMRT cell were diffusion loading, 30-hour movie, 2-hour immobilisation time, 4-hour pre-extension time, 60-80pM on plate loading concentration. We generated 11.7 million PacBio HiFi reads (227 Gb of sequence), corresponding to a haploid genome coverage of 32.9x. The average HiFi read length was 19.4 Kbp.

Dovetail Omni-C library preparation and sequencing

Sample material for the Omni-C library preparation was 300mg of the youngest leaf and meristem material from plants dark adapted for 5 days then flash frozen at harvest. The Omni-C library was prepared using the Dovetail® Omni-C® Kit (SKU: 21005) according to the manufacturer’s protocol²⁶.

Briefly, the chromatin was fixed with disuccinimidyl glutarate (DSG) and formaldehyde in the nucleus. The cross-linked chromatin sample was then digested in situ with 0.05µl DNAse I. Following digestion, the cells were lysed with SDS to extract the chromatin fragments and the chromatin fragments were bound to Chromatin Capture Beads. Next, the chromatin ends were repaired and ligated to a biotinylated bridge adapter followed by proximity ligation of adapter-containing ends. After proximity ligation, the crosslinks were reversed, the associated proteins were degraded, and the DNA was purified before conversion into a sequencing library (NEBNext® Ultra™ II DNA Library Prep Kit for Illumina® (E7645)) using Illumina-compatible adaptors (NEBNext® Multiplex Oligos for Illumina® (Index Primers Set 1) (E7335)). Biotin-containing fragments were isolated using streptavidin beads prior to PCR amplification. The Omni-C library was quantified by qPCR using a Kapa Library Quantification Kit (Roche Diagnostics 7960204001). The pool was diluted to 2nM and denatured using 2N NaOH before diluting to 20pM with Illumina HT1 buffer. The denatured pool was loaded on an Illumina MiSeq Sequencer for quality control with a 300 cycle MiSeq Reagent Kit v2 (Illumina MS-102-2002) at 10pM concentration with a 1% phiX control v3 spike (Illumina FC-110-3001). The MiSeq was run using control software version 4.0, RTA v1.18.54.430, and the data was demultiplexed and converted to fastq using bcl2fastq2. Following quality control analysis, the library was sequenced on one lane of a 300 cycle NovaSeq X Series 10B Reagent Kit (Illumina 20085594). For this run, the library was diluted down to 0.75nM using EB (10mM Tris pH8.0) in a volume of 40µl before spiking in 1% Illumina phiX Control v3. This was denatured by adding 10µl 0.2N NaOH and incubating at room temperature for 5 mins, after which it was neutralised by adding 150µl of Illumina’s preload buffer, of which 160µl was loaded onto the NovaSeq X Plus for sequencing. The NovaSeq X Plus was run using control software version 1.0.1.7385 and was set up to sequence 150bp paired-end reads. The data was demultiplexed and converted to fastq using bcl2fastq2. We generated 1.28 billion reads representing 56.7x haploid genome coverage.

Illumina RNA-Seq library preparation and sequencing

Five root samples and five leaf samples were used to generate RNA-Seq libraries. This was performed on the Perkin Elmer (formerly Caliper LS) Sciclone G3 (PerkinElmer PN: CLS145321) using the NEBNext Ultra II RNA Library prep for Illumina kit (NEB#E7760L) NEBNext Poly(A) mRNA Magnetic Isolation Module (NEB#E7490L) and NEBNext Multiplex Oligos for Illumina® (96 Unique Dual Index Primer Pairs) (E6440S/L) at a concentration of 10µM. One µg of RNA was purified to extract mRNA with a Poly(A) mRNA Magnetic Isolation Module. Isolated mRNA was then fragmented for 12 minutes at 94°C, and converted to cDNA. NEBNext Adaptors were ligated to end-repaired, dA-tailed DNA. The ligated products were subjected to a bead-based purification using Beckman Coulter AMPure XP beads (A63882) to remove unligated adaptors. Adaptor Ligated DNA was then enriched by 10 cycles of PCR (30 secs at 98°C, 10 cycles of: 10 secs at 98°C _75 secs at 65°C _5 mins at 65°C, final hold at 4°C). The size of the resulting libraries was determined using a Perkin Elmer DNA High Sensitivity Reagent Kit (CLS760672) with DNA 1K / 12K / HiSensitivity Assay LabChip (760517) and the concentration measured with a Quant-iT™ Reader) assay from ThermoFisher (Q-33120). The final libraries were pooled equimolarly and quantified by qPCR using a Kapa Library Quantification Kit (Roche Diagnostics 7960204001).

The pool was diluted down to 0.5nM using EB (10mM Tris pH8.0) in a volume of 18µl before spiking in 1% Illumina phiX Control v3. This was denatured by adding 4µl 0.2N NaOH and incubating at room temperature for 8 mins, after which it was neutralised by adding 5µl 400mM tris pH 8.0. A master mix of DPX1, DPX2, and DPX3 from Illumina’s Xp 2-lane kit was made and 63µl added to the denatured pool leaving 90µl at a concentration of 100pM. This was loaded onto a single lane of the NovaSeq SP flow cell using the NovaSeq Xp Flow Cell Dock before loading onto the NovaSeq 6000. The NovaSeq was run using NVCS v1.7.5 and RTA v3.4.4 and was set up to sequence 150bp PE reads. The data was demultiplexed and converted to fastq using bcl2fastq2. A total of 253 million and 234 million reads were generated for root and shoot samples respectively.

PacBio Iso-Seq library preparation and sequencing

Iso-Seq libraries were generated for root and shoot samples. The five shoot extractions were pooled for the shoot library and a single root sample was used for root. The libraries were constructed starting from 356-482ng of total RNA per sample. Reverse transcription cDNA synthesis was performed using NEBNext® Single Cell/Low Input cDNA Synthesis & Amplification Module (NEB, E6421). Each cDNA sample was amplified with barcoded primers for a total of 12 cycles. Each library was prepared according to the guidelines laid out in the Iso-Seq protocol version 02 (PacBio, 101-763-800) using SMRTbell express template prep kit 2.0 (PacBio, 102-088-900). The library pool was quantified using a Qubit Fluorometer 3.0 (Invitrogen) and sized using the Bioanalyzer HS DNA chip (Agilent Technologies, Inc.).

The loading calculations for each Iso-Seq library were calculated using the PacBio SMRTlink Binding Calculator v.10.2.0.133424 and prepared for sequencing according to the library type. Sequencing primer v4 was annealed to the Iso-Seq library pool and complexed to the sequencing polymerase with the Sequel II binding kit v2.1 (PacBio, 101-843-000). Calculations for primer to template and polymerase to template binding ratios were kept at default values for the library type. Sequencing internal control complex 1.0 (PacBio, 101-717-600) was spiked into the final complex preparation at a standard concentration before sequencing for all preparations. The sequencing chemistry used was Sequel® II Sequencing Plate 2.0 (PacBio®, 101-820-200) and the Instrument Control Software v10.1.0.125432.

Each Iso-Seq library was sequenced on the Sequel IIe instrument with one Sequel II SMRT®cell 8M cell per library. The parameters for sequencing per SMRTcell were diffusion loading, 30-hour movie, 2-hour immobilisation time, 2-hour pre-extension time, 60pM on plate loading concentration.

We generated 3.9 million CSS reads for the root sample and 3.8 million CCS reads for the shoot sample. These were filtered to identify Full-Length Non-Concatamer (FLNC) reads resulting in 3.6 and 3.4 million FLNC reads for root and shoot, respectively. Clustering these individually generated 258,543 transcripts for root and 212,332 transcripts for shoot. In addition, root and shoot FLNC reads were combined and clustered to generate a set of 410,286 transcripts.

Genome survey

A previous study reported that A. aequalis has seven chromosomes²⁷. Prior to genome sequencing, we first estimated genome size using flow cytometry for four individual plants of A. aequalis line 828217 which gave a mean estimated genome size of 3.45 Gb/1C (Table 1).

We used FastK²⁹ (v1.1) to count k-mers (k=31) in the HiFi reads, then GenomeScope.FK (based on GenomeScope 2.0³⁰) to explore genome characteristics. In order to estimate genome size, the haploid k-mer coverage was set to 33.

FastK -k31 -T16 -M200 -P. -Ntmp_fastk reads.fastq.gz
Histex -G reads.hist > hist_all.out
GeneScopeFK.R -i hist_all.out -o all_out -k 31 --kmercov 33

The estimated genome size from GenomeScope was 2.9 Gb (Figure 2), which is 0.55 Gb less than the flow cytometry estimate of 3.45 Gb. However, genome size estimates using k-mers are often smaller than those obtained by flow cytometry due to the challenges of assembling the repetitive fraction of the genome. Computational methods try to avoid overestimating genome size by ignoring high frequency k-mers which come from a mix of real repeats and artifacts such as organelles or small contaminants. GenomeScope estimated the heterozygous percentage of the genome as 0.058% and the k-mer profile shows a large single peak centered around coverage 73 representing the homozygous part of the genome. There is a slight deviation from the model on the left side of the peak (marked with a red arrow) representing the heterozygous part of the genome. This indicates the genome is less heterozygous than expected for an outcrossing species.

Open in new tabFigure 2: GenomeScope profile of Alopecurus aequalis HiFi reads using GenomeScope.FK.

K-mer count distribution (in blue) from the HiFi reads showing a single homozygous peak at a coverage of 73, and a peak representing repeats at approximate coverage 150. The summary above the plot shows the estimated genome size in bp (len), unique fraction (uniq), homozygous fraction (aa), heterozygous fraction (ab), k-mer coverage of the heterozygous peak (kcov), PCR error estimation (err), PCR duplication estimation (dup), k-mer length (k), and ploidy (p). The red arrow indicates where the observed deviates from the model representing the heterozygous part of the genome.

Genome assembly

Hifiasm³¹ v0.16.1 was used to generate a contig assembly and the homozygous coverage was specified as 73, corresponding to the position of the homozygous peak in the GenomeScope plot generated from the reads (Figure 1). Hifiasm generates assembly graphs for primary contigs and haplotype 1 and 2.

hifiasm -o fx_run1.asm -t 64 --hom-cov 73 0.5_hifi_reads.fastq

Reads assembled into 1,034 primary contigs with a total size of 2.9 Gb and a contig-N50 of 374.7 Mb (Table 2). The assembly size was similar to the k-mer based estimate and smaller than the flow-cytometry estimate. The total size of the individual haplotypes was similar to the primary contigs but the contiguity was lower (348 Mb and 244 Mb for haplotypes 1 and 2 respectively).

The contiguity of this contig assembly was exceptionally high with the 13 longest contigs comprising 98% of the assembly. To determine whether these contigs represented whole chromosome arms we aligned contigs to the assembly of A. myosuroides¹⁸ and H. vulgare³² ‘Morex’ V3. Ten contigs were identified that represented chromosome or chromosome-arm level sequences. We also found two contigs with telomere sequences (TTTAGGG) on both ends and seven with telomere sequences on one end indicating that we had assembled chromosome or chromosome-arm level sequences. The low heterozygosity in A. aequalis estimated by GenomeScope was confirmed by identifying single-copy k-mers from a comparison of both haplotypes to the reads (Figure 3).

Open in new tabFigure 3: Content differences between the two assembled haplotypes of Alopecurus aequalis.

K-mer copy-number spectra from KAT³³ comparing both haplotypes to the reads. The majority of the kmers are present twice (once in each haplotype: purple peak), the red single-copy region represents kmers that exist in a single haplotype.

Contig scaffolding was performed using the Illumina paired-end Omni-C reads. The Arima Genomics Hi-C mapping pipeline³⁴ was used to map the reads to the contigs before scaffolding with YaHS³⁵. The mapping pipeline uses BWA³⁶ (v.0.7.12) to map reads 1 and 2 separately before combining them into a single BAM file which is then deduplicated using Picardtools³⁷ MarkDuplicates (v.2.25.7). The deduplicated BAM file was used as input to YaHS. This generated seven chromosome-level scaffolds, in agreement with previous cytogenetic studies²⁷, and 421 sequences not incorporated into scaffolds.

To validate the scaffolding, Omni-C reads were mapped back to the scaffolds using BWA and the resulting contact map visualised using PretextView³⁸. This identified one scaffold as retained haplotypic duplication and indicated a potential join between two larger scaffolds (scaffold_7 and scaffold_8). It also highlighted that many of the smaller scaffolds represented contamination. We removed the scaffold representing haplotypic duplication and joined scaffolds 7 and 8 into a larger scaffold. The final contact map is shown in Figure 4.

Open in new tabFigure 4: The PretextView contact map for Alopecurus aequalis after scaffolding and curation.

Seven chromosomes in descending size order are represented by the diagonal red line with centromeric repeats showing as green regions. This map was generated after removal of one scaffold identified as a haplotypic duplication and the merging of scaffold_7 and scaffold_8.

Contaminated short sequences were identified and removed using NCBI BLAST+³⁹ megablast (v2.12.0). The final assembly contained seven chromosome-level sequences and 145 shorter sequences not included in scaffolds. The chromosome-level scaffolds represented 2.8 Gb of sequence (99.5% of the assembly) and the total size of the 145 unassigned scaffolds was 15 Mb. Chromosome-level scaffolds were labeled in descending size order (Table 3).

Genome annotation

Annotation was performed on contigs due to the high contiguity of the assembly at this stage. The annotation was lifted over to the scaffolds once they had been generated.

Genome overview

The A. aequalis genome contains seven chromosomes ranging in size from 521 to 339 Mb (Figure 5). Gene density is highest at the distal ends of the chromosomes with very few genes in the centromeric regions. The distribution of transposable elements is quite even over the length of the chromosomes. The distribution of Ty3/Gypsy LTR retrotransposons (LTR RTs) is highest in gene-poor regions with fewer found in the gene-rich distal ends of the chromosomes. Ty1/Copia LTR RTs are distributed throughout the chromosomes with fewer in centromeric regions.

Open in new tabFigure 5: Overview of the Alopecurus aequalis genome.

Distribution of high-confidence protein-coding genes (blue), distribution of transposable elements (green), distribution of Ty3/Gypsy long-terminal repeats retrotransposons (LTR-RTs) (red) and distribution of Ty1/Copia LTR RTs (pink).

Comparative genomics analysis

GENESPACE⁶⁷ was used to assess synteny between A. aequalis, A. myosuroides¹⁸ and Hordeum vulgare cultivar ‘Morex’⁶⁸ to identify large structural rearrangements. Although a second A. myosuroides genome is available¹⁹, our analyses showed no differences between the two versions. Therefore, we used the genome from Cai et al.¹⁸ in all of our analyses.

The seven chromosomes of A. aequalis are more similar in structure to H. vulgare (Figure 6A). Six of the seven chromosomes show a high level of synteny to H. vulgare with A. aequalis chromosome 1 comprising two syntenic blocks from H. vulgare chromosomes 4H and 5H. There is also evidence of a small translocation between A. aequalis chromosome 6 to H. vulgare chromosomes 4H.

Open in new tabFigure 6: Whole genome comparison between Hordeum vulgare, Alopecurus aequalis and Alopecurus myosuroides.

A: GENESPACE comparison showing synteny between the 3 species (* indicates that the chromosome has been reversed). B: Detailed alignment showing the position of the A. aequalis chromosome 1 centromere (in red) in relation to the breakpoint.

Five chromosomes of A. aequalis show a high level of synteny to A. myosuroides. The chromosome arms of A. aequalis chromosome 1 are syntenic to two regions of A. myosuroides chromosomes 1 and 6. The break in synteny appears to occur in the centromeric region of A. aequalis chromosome 1, estimated to be between 220 and 290 Mb according to the Hi-C contact map (Figure 4) which also corresponds to the drop in gene density in this region of the chromosome (Figure 5). More detailed alignments show A. aequalis chromosome 1 aligns to A. myosuroides chromosome 1 from 0-240 Mb and to A. myosuroides chromosome 6 from 300-521 Mb (Figure 6B).

Chromosome 4 of A. aequalis is syntenic to two large regions on A. myosuroides chromosomes 1 and 6. This relationship is more complex, showing several internal rearrangements within the larger syntenic region. It should be noted that the differences between A. aequalis chromosomes 1 and 4 compared to A. myosuroides chromosomes 1 and 6 were evident in the contig stage of assembly.

Genes involved in herbicide resistance

Cytochrome P450 genes have been identified as the main genes involved in non-target site herbicide resistance^6,7. In A. myosuroides¹⁸, where the P450 gene family has significantly expanded compared to Arabidopsis and rice, 506 loci were identified, mainly on chromosomes 2 and 3. The functional annotation of A. aequalis identified 513 P450 genes, suggesting this species has also experienced an expansion of genes involved in non-target site herbicide resistance (Figure 7).

Open in new tabFigure 7: Locations of cytochrome P450 gene loci in Alopecurus aequalis.

The number in brackets shows the total number of loci on that chromosome. Gene identifiers are shown without the Alaeq_EIv0.2_ prefix for clarity.

Data Records

All read datasets used in the assembly and annotation of Alopecurus aequalis are available at the National Center for Biotechnology Information (NCBI) under Project ID PRJEB75647. Genome assembly and annotation files are available from https://opendata.earlham.ac.uk/opendata/data/wright_sep2024_orange_foxtail/.