The code in this repository is used for the primary analysis presented in the paper Large language model-based biological age prediction in large-scale populations (Nature Medicine, 2025; https://doi.org/10.1038/s41591-025-03856-8). The repository consists of two main components: vllm_inference and Statistical_Analysis.
1.vllm_inference: This part contains the inference code that utilizes large language models to predict multidimensional aging for individuals. The inference is implemented using the vllm framework.
2.Statistical_Analysis: This component focuses on the validation and application of biological age predictions generated by the large language models. It mainly uses R for subsequent statistical analysis.
For an overview of this study, see the figure below.
