The Genomics for Australian Plants (GAP) framework initiative – developing genomic resources for understanding the evolution and conservation of the Australian flora

Although there has been a marked acceleration in the availability of plant nuclear genomic resources in recent years, most of these are for crop species and wild relatives, and forest trees. Thus, resources are very unevenly distributed across the plant tree of life, hampering use to understand patterns of genomic diversity and processes that drive genome evolution. The GAP Reference Genomes stream aimed to address this resource bias for Australian lineages of flowering plants (angiosperms). Initially, a Working Group was formed to consider: (1) which of the increasing array of short- and long-read genomic sequencing technologies are most appropriate for the task; and (2) how to prioritise taxa for sequencing.

To identify orders and families in the Australian angiosperm flora that lacked genomic resources, the GAP Steering Committee led the compilation of a list of all plant nuclear genomic resources available from public repositories (NCBI genomes, NCBI Sequence Read Archive, PlaBiPD), including completed genomes, draft genomes or assemblies and other sequence data (e.g. RNAseq and genome skimming). This list is available online (see links under GAP Data and Resources Availability section below). Using this list for priority setting, community participation in the GAP Reference Genomes stream was invited through Requests for Partnership in four phases. This approach enabled the consortium to adapt the program in response to new technological applications as these became available and progressively tackle larger genomes. Plants with diploid genomes up to 2 gigabases (Gb; 1C-value) were targeted as these were considered tractable for sequencing to sufficient depths for assembly. Additionally, the lower cost of sequencing for genomes in this size range compared with larger genomes meant that the GAP Framework Initiative could sequence more representatives to fill taxonomic gaps across the tree of life.

Extracting high molecular weight DNA suitable for reference genome assembly was a significant challenge for nearly all taxa and exceptionally difficult for some (Jones et al. 2021). To address this difficulty, members of the consortium developed dedicated protocols suitable for extracting high molecular weight DNA for long read sequencing of diverse non-model plants. This included virtual engagement with the wider scientific community for interactive protocol development, discussion and Open Access sharing through Protocols.io (Teytelman et al. 2016). Nevertheless, several projects could not be completed because obtaining DNA of sufficient quality and quantity proved intractable. All of the species for which nuclear genomes were sequenced through the GAP Framework Initiative are listed in Table 1, including those that are not yet completed at time of publication.

To date, the GAP Reference Genomes stream has delivered: (1) new protocols for high-molecular weight DNA extraction, clean-up and size selection for long read sequencing (Jones et al. 2021; Jones and Schwessinger 2021); (2) an approach to base-caller selection to improve actual accuracy of nanopore sequencing (Ferguson et al. 2022); (3) case studies in the use of long read assemblies to reveal structural diversity in organellar genomes (Syme et al. 2021); (4) genome size estimates for 10 species (Chen et al. 2022; McLay et al. 2022; Schmidt-Lebuhn and Cantrill 2023; T. McLay, unpubl. data; A. Schmidt-Lebuhn and R. Fowler, unpubl. data); (5) an online concatenated database of published genomic resources compiled from public repositories; and (6) published genomic resources for Australia’s national floral emblem, the golden wattle, Acacia pycnantha Benth. (Syme et al. 2021; McLay et al. 2022) and the state floral emblem of New South Wales, the waratah, Telopea speciosissima (Sm.) R.Br. (Chen et al. 2022).

The status of all 35 reference genome projects is provided in Table 1. Together, these 35 projects provide genomic resources for 24 orders and 30 families, and no published genomic resources exist for 4 of these orders (Austrobaileyales Takht. ex Reveal, Berberidopsidales Doweld, Dilleniales DC. ex Bercht. & J.Presl, Paracryphiales Takht. ex Reveal) and 10 of these families (Austrobaileyaceae Croizat, Berberidopsidaceae Takht., Cunoniaceae R.Br., Dasypogonaceae Dumort., Doryanthaceae R.Dahlgren & Clifford, Dilleniaceae Salisb., Eupomatiaceae Orb., Monimiaceae Juss., Paracryphiaceae Airy Shaw, Winteraceae R.Br. ex Lindl.). Most of the genome projects that are nearing completion will be published in a single overarching and descriptive GAP genome paper.

An understanding of phylogenetic relationships underpins all evolutionary biology, including systematics. To date, plant molecular phylogenetic research in Australia has been largely directed by individual researcher or research institution needs and interests, rather than as a nationwide coordinated effort, with different approaches to marker choice and sampling across groups limiting the interoperability and reusability of data. The GAP Phylogenomics activity stream aimed to transform phylogenetic research on the Australian flora by producing a sequence data resource of common markers and a phylogenetic tree containing at least 95% of Australian native angiosperm genera (i.e. 2037 of 2144 total genera, APC, see https://biodiversity.org.au/nsl/services/, accessed 28 June 2024) – the AAToL. Progress has been achieved by supporting and coordinating collaborators in achieving research goals where contributions were made to building the AAToL.

Two working groups were formed to design and coordinate the GAP phylogenomics component. The Phylogenomics Working Group conceived the project aims and design, engaged the community and coordinated the sampling strategy. The Phylogenomics Bioinformatics Working Group provided guidance and recommendations for approaches to data analysis, development of bioinformatic tools for analysis of target sequence capture datasets and generation of the AAToL tree, and provided bioinformatics support.

The Phylogenomics Working Group advised the adoption of target capture (also known as hybrid enrichment, target enrichment and exon capture) as the sequencing approach that offers reliable retrieval of hundreds or thousands of loci from across the nuclear genome (Cronn et al. 2012). A decision was subsequently made to engage with the Plant and Fungal Trees of Life (PAFTOL; Baker et al. 2022) project led by the Royal Botanic Gardens Kew since 2015 that aims to build the tree of life for all angiosperm genera. Given the aligned approach, GAP and PAFTOL agreed to work in close partnership toward resolving the AAToL, with the GAP team reciprocating by making a subset of data available to the PAFTOL project (Zuntini et al. 2024). This has been achieved by coordinating sampling and methods for genomic data generation, and collaboration on scientific research outputs.

The aim was to construct the AAToL to genus level in two stages with community participation invited through Requests for Partnership:

The Phylogenomics Bioinformatics Working Group generated several key bioinformatics resources including: (1) a set of recommended approaches and software to use for preparation and analyses of target enrichment data for consortium-scale phylogenomics projects (link under the GAP Data and Resources Availability section below); (2) updated scripts for the recommended software (Yang and Smith 2014; Johnson et al. 2016) to provide novel options, remove bugs and seamlessly use the outputs of the former as inputs for the latter; (3) and a containerised analysis workflow for HybPiper (ver. 2.0, see https://github.com/mossmatters/HybPiper) and ParaGone (ver. 1.1, see https://github.com/chrisjackson-pellicle/ParaGone) (Jackson et al. 2023) to overcome challenges with software compatibility, unfamiliarity with relevant programming languages and the complexity involved in running numerous analysis steps required to use the recommended software (Jackson et al. 2023); (4) Python packages of HybPiper and ParaGone (links under the GAP Data and Resources Availability section below). Additionally, work on AAToL led to two methods and resources studies. The first study, undertaken in partnership with PAFTOL researchers, developed a new target file – ‘mega353’ – for recovery of sequences of Angiosperms353 loci from target capture data sets (McLay et al. 2021). The use of this new target file, rather than the original target file of Johnson et al. (2019), can substantially increase the percentage of on-target reads, locus recovery at 75% length and the total length of the concatenated loci compared to the default Angiosperms353 file. The second study explored conflict in a range of AAToL Stage 2 datasets and developed guidelines for researchers to detect, characterise, manage and interpret conflict in target capture datasets (Joyce et al. 2024, 2025).

Together, the pilot and Stage 1 of the AAToL generated Angiosperms353 sequence data for 2204 samples representing 2019 genera. Sequence data for 650 of these samples were provided by the PAFTOL project and data for 75 samples were sourced from genomic repositories such as the NCBI Sequence Read Archive (see https://www.ncbi.nlm.nih.gov/sra). A total of 2001 samples passed quality control, resulting in a draft phylogenomic tree including 1927 (90%) of 2144 native Australian angiosperm genera. Sequence data are available to the GAP consortium through the Bioplatforms Australia Data Portal. In a reciprocal data sharing arrangement with PAFTOL, data for 770 samples sequenced through the GAP Framework Initiative were made publicly accessible from the Kew Tree of Life Explorer (treeoflife.kew.org; Baker et al. 2022) that currently provides Angiosperms353 data for >10,000 species (data release 3.0, April 2023). AAToL contributed 763 samples (~8%) to a comprehensive analysis of the angiosperm tree of life including nearly 8000 (~60%) angiosperm genera that revealed new insights into angiosperm origins and diversification (Zuntini et al. 2024). The data have also contributed to several other publications produced by PAFTOL collaborators (e.g. Joyce et al. 2023; Elliott et al. 2024; Pérez-Escobar et al. 2024; Grass Phylogeny Working Group III 2025; Helmstetter et al. 2025). Samples released though the Kew Tree of Life Explorer are also publicly available from the Sequence Read Archive (European Nucleotide Archive – ENA – and National Centre for Biotechnology Information – NCBI; project PRJEB49212). The remaining data from Stage 1 of the AAToL will be made publicly available when the comprehensive phylogenetic analysis is completed and published.

The partnership with PAFTOL has greatly facilitated the achievement of GAP aims, most notably by (1) increasing the rate of progress toward completion of the AAToL; (2) expanding the professional networks of Australian scientists, particularly early to mid-career researchers through collaboration with a global team of leaders in plant phylogenomics; and (3) maximising the potential for future re-use of GAP data through adoption of a universal target capture bait set and contribution to global datasets (Baker et al. 2022) and analyses (e.g. Zuntini et al. 2024). Taken together these outcomes substantially advance the capacity for genomics data generation and use by the Australian herbarium community and enable a step change in our understanding of the relationships and evolution of Australia’s flora. In turn, GAP’s engagement has significantly helped expand the reach of PAFTOL and has functioned as a model for the collaborative, open data mindset that PAFTOL aims to foster.

AAToL Stage 2 generated Angiosperms353 sequence data for 10 species complexes (Table 2). In addition, sequence data were generated using OZbaits (Waycott et al. 2021) or SaliBaits (A. Žerdoner Čalasan, pers. comm.) kits for several groups in which those kits were known, or expected, to outperform Angiosperms353 (Table 2). Data analysis, including an assessment of the utility of each bait set and preparation of outputs will be reported in manuscripts focused on each taxonomic group. Papers on Chamelaucieae (Nge et al. 2025), Pultenaea Sm. (Barrett et al. 2024) and Minuria DC. (Schmidt-Lebuhn et al. 2024) have been published and manuscripts on Drosera L., Hibbertia Andrews, Santalaceae, Teucrium L. and on navigating conflict in target capture datasets have been submitted for this Special Collection (Table 2). Several other studies utilising AAToL Stage 1 and Stage 2 data have also been published (e.g. on Aglaia Lour., Cooper et al. 2023; Mirbelieae, Barrett et al. 2024; Pogonolepis Steetz, Schmidt-Lebuhn 2022; and Senecio L., Schmidt-Lebuhn et al. 2022).

The data and lessons from Stages 1 and 2 have enabled the community to commence planning for a continuation of the project beyond the GAP Framework Initiative: AAToL Stage 3. Stage 3 aims to resolve the AAToL to species level, complete for >95% of accepted Australian species. Avenues for resourcing this large undertaking (>18,000 species) and approaches for completion are currently being explored.

The Conservation Genomics stream aimed to provide genomic information to support conservation of the Australian flora. Conservation genomics covers a range of activities according to the management questions being addressed, including resolution of closely related taxa and species complexes to identify taxa of conservation concern; conservation units based on evolutionary lineages; genomic diversity to inform conservation interventions (e.g. augmentation, germplasm capture, translocation and restoration); hybridity and hybrid origin of species; mating system to understand level and pattern of outcrossing; and genotype-environment association analysis to identify signals of adaptation across climate gradients.

Studies among these aiming to resolve species complexes containing suspected taxa of conservation concern were deemed the most suitable for support through the GAP Framework Initiative. The Conservation Genomics Working Group was established to identify species complexes most in need of resolution due to conservation concerns and taxonomic complexity, and to engage the research community in identifying groups in which relevant taxonomic expertise could take advantage of the genomic input and data. The Conservation Working Group decided to take a specific molecular approach (double digest RADseq–ddRAD; Peterson et al. 2012) to create interoperable data sets and develop capability for population level data analyses. Projects to be supported were identified using the following criteria:

A Request for Partnership elicited 15 responses involving 25 institutions, including 7 state and institutional herbaria (Table 3). The Steering Committee was keen to stimulate collaboration among partners and the taxonomic and genomics communities, and assisted by linking up collaborators in cases in which proposals lacked key capability on the team (e.g. genomic data analysis expertise). Some projects did not proceed due to issues with DNA sample quality, particularly for projects that exclusively utilised herbarium samples. For four projects, the same samples were also sequenced using the Angiosperms353 bait kit to evaluate the comparative performance of a target capture approach for the resolution of species boundaries within species complexes.

Ten of the fifteen projects proposed provided DNA of suitable quality for ddRAD sequencing whereas five projects, on Allocasuarina L.A.S.Johnson, Lepidosperma Labill. (x 2), Melichrus R.Br. and Thelymitra J.R.Forst. & G.Forst. did not proceed (Table 3). For most projects, herbarium specimens did not yield DNA of usable quality for ddRAD analysis and new collections were required to provide fresh material. The workflow for these projects included the assembly of sequence libraries and SNP calling using an optimised ipyrad (Eaton and Overcast 2020) approach, followed by custom filtering to provide high quality SNP data sets for downstream analyses. These analyses employed a combination of phylogenetic and population genetic approaches to obtain information on the genomic variation within and among the sampled populations of each species complex. This included concatenation and coalescent phylogenetic analyses, and ordination, admixture and distance-based population genetic analyses.

For each of the species complexes evaluated, the genetic results were interpreted in conjunction with morphological data to refine taxonomic hypotheses of relationship and identify undescribed, sometimes cryptic, taxa or hybrid populations – results that will directly underpin conservation assessments and management actions. Furthermore, the results provided insight into the morphological complexity within each group and the informative characters for species identification. Results for projects on Geleznowia Turcz., Isopogon R.Br. ex Knight and Zieria Sm. have been published (Anderson et al. 2023, 2024; Orel et al. 2023a, 2023b), and results for Gompholobium Sm. are included in this Special Collection (Simmons et al. 2025).

The GAP Framework Initiative aimed to build genomic capacity across Australian botanic gardens and herbaria, and provide tools to enable genomic data to be used to identify, classify and conserve biodiversity. Training and capacity building across the GAP consortium has been an inherent and integral part of the development and deployment of workflows undertaken at all stages of the reference genomes, phylogenomics and conservation genomics projects. This included sampling of curated collections for genomic analysis, sample metadata curation, wet lab sample preparation for genomic analysis, bioinformatic analysis of genomic data and preparation of manuscripts. The Wet Lab and Bioinformatics Working Groups reviewed current methods and available resources to develop best practice approaches and protocols, tools and training resources for GAP activities. Resources were also developed to train researchers in the use of the containerised bioinformatic pipelines (see below under the GAP Data and Resources Availability).

A webinar and workshop series for the analysis of target capture sequence data were delivered to 20 participants at the virtual Australasian Systematic Botany Society conference in July 2021, and the workshop materials and webinar recordings have been made publicly available (see below). Additionally, in response to key challenges identified in the early stages of the GAP Framework Initiative, resources were developed and made accessible to encourage and support best practice approaches for plant genomic initiatives and ongoing capacity building across the sector. These resources include: (a) protocols for the extraction of high molecular weight DNA for long read sequencing; (b) analysis guides for target sequence capture projects; (c) containerised bioinformatic software pipelines for the analysis of target sequence capture data (Jackson et al. 2023); and (d) scripts for preparing target sequence data for upload to ENA. Links for the first three of the above public resources are provided in the GAP Data and Resources Availability section, below. The last will be made available on a suitable Open Access platform in the near future.

In addition to the specific, measurable outputs detailed above, GAP has generated many intangible benefits for the community. A wide range of plant scientists, students and collections staff in herbaria, universities and botanic gardens – more than 210 people in total – was engaged in GAP activities. Mid-career and established researchers who have utilised legacy molecular approaches, such as Sanger sequencing or ISSR/SSR, for large parts of their careers have experienced a step change in their capacity to generate and interpret genome scale datasets. Staff of smaller institutions that lack capacity for genetic research of any kind have enjoyed unprecedented opportunities to participate in large genomics projects. Postgraduate students and early career researchers already well versed in genomic analysis have been able to undertake specific training in areas of need and access large volumes of data that would have been difficult to finance otherwise. All participants have benefited from opportunities to engage and collaborate with leading researchers globally, and through these collaborations have developed a greater understanding of the applications of genomic technological applications and data not only for evolution and systematics research, but also for collections development and management. These opportunities, critically enabled by the consortium model (Lowman et al. 2024) have undoubtedly led to higher-quality and more effective research outputs, and expanded professional networks, both of which are likely to have a positive influence on careers.

The GAP Framework Initiative consortium model has extended and strengthened collaborative networks across the sector nationally and internationally. Researchers and technical staff from the plant collections, systematics, genomics and bioinformatics communities were represented from the initial consultations undertaken to develop the consortium through to the assembly of the Steering Committee and advisory Working Groups. The wider consortium engagement across all GAP projects has strengthened cross-institutional collaboration and awareness on collections curation, wet lab and bioinformatics activities, and promoted coordination of approaches and activities to maximise efficiency and the potential for data sharing and re-use. Beyond Australia, GAP researchers have collaborated with hundreds of scientists across the global plant systematics and genomics community, principally through involvement in the PAFTOL consortium. These collaborations have produced a step change in global angiosperm phylogenomics and high-impact publications (e.g. Larridon et al. 2021a; McLay et al. 2021; Zuntini et al. 2024), and have sparked a large, diverse range of new projects with researchers worldwide (e.g. Maurin et al. 2023; Elliott et al. 2024; Saunders et al. 2024; Grass Phylogeny Working Group III 2025; Helmstetter et al. 2025). In turn, these relationships have influenced global thinking and practice, for example through the endorsement and uptake of universal target capture bait kits (e.g. McDonnell et al. 2021; Fonseca et al. 2024), and publication of new taxonomic frameworks (e.g. Larridon et al. 2021b; Pillon et al. 2024).

To date, the GAP Framework Initiative has generated reference genome sequence data for 35 species, including the first genomic resources for 10 families and 4 orders, target sequence capture datasets representing 87% of Australia’s native flowering plant genera and reduced representation genomic datasets for 10 species complexes of conservation concern. These outputs have considerably increased taxonomic coverage of genomic resources for Australian plants. All raw data are immediately released to GAP consortium partners on the Bioplatforms Australia Data Portal and made accessible on public genomic repositories upon publication of data papers. Through mandating that source material must be vouchered in an Australian herbarium, the GAP Framework Initiative has enhanced curated collections through the extended specimen concept (Webster 2017; Lendemer et al. 2020) by enriching physical collections with genomic data.

The GAP consortium includes curators, researchers, technicians and other staff from more than 30 institutions across Australia and internationally. Consortium members have benefited from this collaborative program of work through involvement in genomic data generation, analysis, interpretation and dissemination. These include sampling and curating metadata from herbarium collections, wet lab preparation, bioinformatics analysis, preparation of research outputs and participation in training workshops. These researchers constitute an upskilled workforce equipped with improved capacity to generate and leverage genomic resources to identify, classify and manage biodiversity, and to improve the management, development and use of collections. Some institutions have already implemented improvements to data and specimen management protocols based on GAP processes and workflows.

The outputs from the GAP Framework Initiative directly contributed to progress against various national science goals, in particular the Decadal Plan for Taxonomy and Biosystematics 2018–2027 (Taxonomy Decadal Plan Working Group 2018) through Strategic action 1.3 ‘…build a comprehensive framework to understand the evolution of the Australian and New Zealand biota’, and Australia’s Strategy for Nature through Objective 10: ‘Increase knowledge about nature to make better decisions’, specifically progress measures 10A (Explicit science and knowledge programs to support effective management of biodiversity) and 10C (Systems capturing data on the diversity of Australian nature and how ecosystems function).

Australia’s biological heritage includes a diverse flora with a complex evolutionary history, widespread and restricted species, both long term natural isolation and more recent fragmentation, and both old and young lineages. The GAP Framework Initiative has helped improve upon Australia’s world-class infrastructure for understanding and managing this heritage by enhancing and enriching knowledge, and human and institutional resources. The reference genomes are improving our understanding of genomic diversity, structure and evolution associated with major Australian plant radiations (e.g. Chen et al. 2022; McLay et al. 2022). The AAToL will provide a common phylogenomic dataset for the flora and ultimately, the phylogenetic backbone for the Flora of Australia project and data portals such as the Atlas of Living Australia. Finally, the conservation genomics projects are resolving the taxonomy and phylogeography of complex groups to enable species documentation and define units for conservation management (e.g. Anderson et al. 2023, 2024).

The GAP Framework Initiative has not only built capacity and delivered significant new data but also revealed where the approaches taken were fit for purpose and where not. For example, a key lesson was the recognition that large collaborative consortium approaches require dedicated human resourcing and robust governance structures. Critical to our success has been the provision of dedicated project management, community engagement and bioinformatics professionals, with Steering and Working groups overseeing and managing each of the major project domains. We learned the importance of converging on shared and scalable computational resources at the national level, while making use of common workflows tailored for GAP Framework Initiative aims and use cases. Another key lesson was understanding that a one-size fits all approach to source material and extraction protocols does not work across the broad range of molecular approaches employed in this project. For example, although herbarium specimens generally yielded DNA suitable for target capture sequencing, these were generally unsuitable for the ddRAD analyses undertaken through conservation genomics projects, such that new, fresh collections were required. For the reference genomes projects, most teams had access to adequate amounts of fresh tissue but still struggled to achieve the high quality, high molecular weight DNA extractions required for long read sequencing. Understanding and developing the most appropriate source material and extraction protocols for each approach was critical to success. Further investment in capacity building around DNA quality will be needed to further realise opportunities for genomics in collections science.

A final key lesson was the value of large-scale collaborative consortia in rapidly increasing capacity and uptake of genomic approaches across a broad scope of research in genome assembly, phylogeny, systematics and conservation genomics. The project developed and strengthened collaborations, and leveraged off the regional or local skills and specific taxonomic expertise across the group. The consortium approach used here provides a potential blueprint for similar initiatives in other countries, and a foundation for further development and application of genomics to plant research in Australia. Future directions for plant biodiversity genomics in Australia are discussed elsewhere in this Special Collection.