Natural language querying of biological databases with large language models

It is attractive to be able to query biological knowledge bases using natural language. Natural language queries are typically translated into structured queries using large language models (LLMs). Here we report the outcomes of a systematic assessment of current practices for natural language querying with LLMs. We find that the best balance between accuracy and flexibility in this context is achieved by multiple LLM agents that can challenge the outputs of each other and interact with a human user. We highlight the need for appropriate benchmarks for the assessment of natural language data mining systems and provide a forward-looking view on how to best structure the application of LLMs to natural language querying.