BRE09 ATLASEA Bot , an AI assistant to explore marine genome diversity

The aim of project is to develop a working prototype of a conversational assistant that allows users to query, using natural language, the data generated as part of a major biodiversity sequencing project ATLASEA . Participants will be provided with a structured dataset, in the form of tabular files and/or a PostgreSQL database containing taxonomic information, sampling metadata, sequencing results and quality indicators. The challenge is to make this data accessible to a wider scientific audience without requiring proficiency in SQL or knowledge of the databases’ internal structure. The chatbot must be able to translate questions formulated in natural language into actionable queries on the data, and then return the results in a comprehensible and well-documented format. Examples of queries include: “How many species have been sequenced in Finistère?”, “Which macroalgae samples were collected in 2024?”, "What are the algae species encoding protein kinases of type GSK?"or “Which taxa have an assembled genome?”. Particular attention will be paid to the traceability of responses, with queries or data sources used clearly displayed. This hackathon will mark the first step towards the development of a scientific assistant specialising in biodiversity, which will eventually be capable of exploring hundred thousands of records and facilitating the use of genomic, taxonomic and ecological data produced by the scientific community.