Perennial ryegrass is the most widely sown forage grass within the UK due to its ability to form productive swards which support grazing-based ruminant agriculture. When used in combination with forage legumes, such as red and white clover, these mixed grasslands can provide both a high energy and protein feed and allow for fewer inputs, due to the nitrogen-fixing properties of the legumes. However, the increasing variability of the climate means that the individual plants which make up these mixed grasslands must be able to persist in the face of a wider range of environmental challenges. Our research is aimed at identifying some of the genetic components of persistence and resilience in these grassland species.

^{Field plots of perennial ryegrass and red clover at Aberystwyth}

Approach: Our approach involves using some of the considerable genetic resources that exist for ryegrasses and clovers to generate novel populations which show interesting variations in trait performances. For ryegrass we are using experiments on individual plants in controlled environments which look at response to water-logging and drought, combined with field based analyses of population performances, i.e., plants grown in dense swards. The results of these trials will indicate genes and genomic regions which may play important roles in conferring persistence and resilience in the face of abiotic stresses. For red clover, we have identified germplasm with contrasting growth habits ranging from erect to prostrate or even creeping, where roots form from nodes of shoots lying on the surface. The erect growth habit is chiefly a result of breeding, while the prostrate habit tends to be associated with natural populations or ecotypes. We believe that growth habit determines how well the plant is able to cope with the selective pressure of grazing or frequent cutting and have generated populations which are derived from crosses between erect versus prostrate growth habit which will allow us to identify key genes in controlling this trait.

Potential impact: Grasslands cover c. 70% of the utilised agricultural area within the UK and, therefore, have a major impact on many aspects of our environment. Because of the extent of these grasslands, incremental improvements in the performance of individual grass or clover plants could generate considerable benefits. For instance, longer-term persistence of mixed grass and legume swards and improved resilience in the face of environmental stresses can result in the need for fewer interventions to maintain or renew the grasslands, thus generating more stable natural environments with lower energy costs.

Key research insights and findings: We have generated populations of ryegrasses and clovers of varying genetic complexities which have been sown as experimental plots in the field. The unusual heat and drought during the summer of 2018 meant that these plots were also exposed to considerable abiotic stress during establishment. Work at the moment is characterising whether the conditions during this field establishment affected the balance of individuals and genes within these populations.

We have evaluated c. 200 individual ryegrass plants taken from genetic resources collections in controlled environments for their responses to drought and water-logging. This has allowed us to identify individuals for further study which will be used in controlled crosses to generate populations for further analysis.

^{Plants being evaluated for water-stress by submersion up to the crown for 3 weeks}

We have conducted a genome-wide association study (GWAS) utilising 716 perennial ryegrass individuals from 90 different locations across Europe grown in the field in Aberystwyth over 2 years. By integrating the results of this study with the publication of a newly developed BAC-based physical map and minimum tiling path genome sequence, we have been able to identify a number of genomic locations associated with variations in flowering time, plant morphology and water soluble carbohydrate accumulation. The latter is a key trait for meeting the energy requirements of grazing animals (doi.org/10.1371/journal.pone.0207412).

We have set up field experiments at Gogerddan and at our upland research site Pwllpeirian using red  clover populations segregating for growth habit traits which are being managed for cutting for silage or simulated grazing. We will use these trials to assess changes in allele frequencies, using genotyping by sequencing on pooled plant material from each plot. We will use the genomic data to investigate how allele frequency changes may be used to identify regions of the genome under selection (Ergon, Skøt et al 2019; Frontiers in Plant Science 10:718. doi.org/10.3389/fpls.2019.00718). We are also collecting phenotypic data from the trials in the form of biomass yield, ground cover and quality, as well as other observational traits. Additionally, we are using a spaced plant trial on which we can make detailed phenotypic observations of the growth habit, and combining this with genotype data using GBS of each plant in the population. We will map traits associated with growth habit to the red clover genome using a GWAS approach. The data from the two experiments will provide us with further insight into this important complex domestication trait, and map genomic regions under selection and underlying the trait. The use of breeder relevant material means that the results will feed into the breeding programme.

^{Overview of red clover experimental plots used in the project. In the foreground are grazed (simulated) plots, and behind them the plots cut  for simulated silage management}

We have undertaken an analysis of gene expression in perennial ryegrass shoots and roots in response to increasing water stress, using 3 different analysis programmes (DeSEQ2, edgeR, limma-voom) to identify differentially expressed genes (DEGs) and comparing associated gene ontology (GO) terms. The most striking findings from this study include that for the shoot: i) while approximately equal numbers of up and down-regulated DEGs were identified at each stage, up-regulated DEGs were more likely to be included in enriched GO terms at earlier stages but down-regulated DEGs more likely to be associated with enriched GO terms at later stages; ii) up-regulated DEGs were often associated with transcription factor and membrane transporter activities and down-regulated DEGs with more general metabolic processes. For the root: i) very few either up- or down-regulated DEGs were identified in earlier stages of drought; ii) at the late stage, heavy-metal associated membrane transporter and cytoskeleton-associated ATPase activities were differentially up- and down-regulated, respectively; v) for both shoot and root, KEGG pathway analysis indicated that ATPase and lignin biosynthesis-associated peroxidase activities were affected by increasing water-stress, doi.org/10.1371/journal.pone.0220518.

^{Inter-sample variation (first two principle components) for gene expression data collected at 35% (blue), 15% (pink), 5% (green) and 1% (purple) soil water content}

We have carried out an analysis of abiotic stress tolerance in a festulolium mapping family (a cross between perennial ryegrass and meadow fescue) which has been characterised both in terms of genetic markers and the relative contribution of the meadow fescue genome. We have discovered genomic associations with response to low phosphorus as well as plant growth and biomass under normal and stressed conditions. An interesting observation is that whole chromosome introgressions of chromosomes 3 and 4 from meadow fescue into the perennial ryegrass background (see image) seem to be particularly associated with plant vigour (doi.org/10.1093/aob/mcy230).

We aim to identify and characterize genes associated with interspecies competition in grassland, and to characterise the genetic and physiological responses of key grassland species to grazing, cutting and other biotic and abiotic stresses.

Approach: We use techniques across different scales including DNA-barcoding, plant neighbour analysis and remote sensing to track the spatial flux of constituent species under different management regimes and across a climatic transect. We are using epigenetic profiling of plants exposed to different stressors in a field environment, and NGS DNA barcoding on faecal samples/offcuts to determine species-specific biomass removal.

Potential impact: We will identify plant responses underpinning resilience to grazing/cutting in the context of competing multi-species swards. In the short term, we will allow improved assembly of commercial species and varietal mixtures tailored to both the growing environment and grazing regime. In the longer term, we will improve forage grass and legume varieties that are more compatible with each other and resilient to grazing/cutting.

Key research insights and findings: Our initial work focussed on developing new approaches to detect changes in sward species composition during establishment or in response to management/climatic variables. We optimised two low cost, high-throughput alternatives to DNA metabarcoding (targeted PCR-RFLP and PCR-HRM) thereby increasing the potential statistical power of sward flux studies. Combined use of all approaches enabled us to process far more samples than by metabarcode-tag-labelling alone. These strategies are now being combined to describe sward dynamics during establishment and will next be used to characterise swards in response to management and across an altitudinal gradient broadly representative of >60% of UK grazing land.

Two contrasting commercial seed mixtures optimised for lowlands and uplands were established across a gradient of the four altitudinal sites. Each site contained five replicated blocks for two contrasting seed mixes (10 blocks in total). Each block was subdivided into sub-blocks (7m x 7m) that were subject to one of four grazing regimes (continual grazing, rotational grazing, simulated grazing and hay cut). Reference plants have been collected from all sites and used to create local reference barcodes suitable for all profiling techniques (core barcodes plus many supplementary loci). To establish species compositions point quadrat samples and pooled point quadrat samples were collected during and then following sward establishment.

	Upland Trial Sites