*Please note that this PhD will be hosted at Royal Botanic Gardens Kew*

Syzygium (Myrtaceae; c 1200 species), is a megadiverse tree genus, encompassing species of economic significance e.g. the clove. Syzygium habitat spans Southeast and Southern Asia, Australia, and the Pacific where species are abundant on acidic, organic-rich soils such as peat swamps, a habitat under strong anthropogenic threat. Despite its environmental and economic importance, taxonomic complexity in Syzygium, lack of taxonomic data, systematic revision and limited evolutionary and ecological understanding mean that Syzygium contributes disproportionally to the ‘taxonomic impediment’, challenging downstream activities such as biodiversity management and conservation.

Mobilisation of existing taxonomic content is achieved through the World Flora Online (WFO), an international initiative. WFO uses a taxonomic backbone to facilitate access to descriptive content for species. Currently species descriptions (in a range of languages, but primarily English) are available for only 38% of accepted species of plants. There is recognised urgency to make descriptions available for all plant species to effectively address the biodiversity crisis currently facing humanity. Only 12% of Syzygium species have descriptions available on WFO, making it an ideal study group for this project. The range of human languages used across the distribution of Syzygium (e.g. French in the Pacific, German in Papua New Guinea), also mean that translation of descriptions will also be explored. WFO plans to extend its descriptive content to include identification tools, in partnership with the PlantNet project.

Literature has long been used as a source of data for the compilation of structured data resources, but these activities were only possible at well-resourced organisations with a comprehensive library. Charles Darwin initiated Index Kewensis (now IPNI) with a legacy to support an editorial team to record bibliographic details of new plant species names. Subsequent projects include the World Checklist of Vascular Plants which scans literature to record taxonomic decisions and information about the distributions of species. These projects continue today.

Digital publishing and literature digitisation programmes like the Biodiversity Heritage Library (BHL) and OpenAlex have built enormous information resources which can be used as a data source, making access more equitable and enabling more participants to engage in literature sourced information extraction. BHL currently includes over 62 million page images,; 3503 of its articles concern Syzygium.

This project will survey the available digitised and born-digital literature for relevant sources to extract basic morphological species descriptions for the large genus Syzygium and to construct identification tools. The project will produce new tools for data extraction from literature delivering key biodiversity informatics and systematic outputs. The key biodiversity informatics output will include new models and gold-standard datasets for the evaluation of information extraction processes. Information extracted via these new approaches will be compared with that mobilised via traditional routes, enabling the community to make informed decisions about how best to populate key informatics resources like the World Flora Online. The key systematic output will be a starting point for a much needed monograph revision at global scale, for the large and taxonomically impenetrable genus Syzygium, with a description made available for every species entry in WFO.

WWF (2024) Living Planet Report 2024 – A System in Peril. WWF, Gland, Switzerland.

Training Opportunities:

Candidates should have (or be about to receive) at least a 2:1 degree in a relevant field, such as computer science, biodiversity informatics, text mining, machine learning, or artificial intelligence. Applicants with a 2:2 degree may be considered if they hold a master’s degree with Distinction or Commendation. The project will suit candidates with a computational background and an interest in applying analytical techniques to biodiversity, or be technically skilled botanists. Training will be provided for methods and analyses specific to the project, ensuring candidates are well-prepared to contribute effectively.

Student profile:

This project will be suitable for students from a wide range of backgrounds. Ideally, they would possess strong computational skills, including experience with machine learning, natural language processing, and programming languages such as R or Python. Students with experience working with large datasets and predictive statistical modelling would be ideal.

How to Apply:

To apply please use the Good Grants system at ai-intervene-dfa.grantplatform.com