2013年6月，厦门大学 化学专业，学士学位；
2019年5月，美国密歇根大学(University of Michigan) 化学专业，博士学位；
2019年2月-2025年1月，百时美施贵宝(Bristol Myers Squibb), post-doc, Senior Scientist, Principal Scientist；
2026年2月至今, 厦门大学生命科学学院，副教授

2013 June, B.S. , Xiamen University, Chemistry;
2019 May, Ph.D. , University of Michigan, Chemistry;
2019 Feb- 2026 Jan, post-doc, Senior Scientist, Principal Scientist, Bristol Myers Squibb;
Associate Professor, School of Life Sciences, Xiamen University, 2026

单分子显微技术（如单分子荧光显微成像）是揭示微观分子过程的变革性工具。通过打破传统系综平均测量的限制，单分子显微技术能够以纳米级空间分辨率和毫秒级时间分辨率，直接观测单个分子的瞬态中间体和罕见动态过程。本课题组致力于将前沿人工智能（AI）与单分子技术深度融合。我们旨在开发通用、高效、自动化的数据分析“元工具”，以克服传统单分子分析中泛化能力差、耗时及主观偏差等瓶颈，进而精准解析复杂生命体系与纳米器件中的瞬态动力学机制。构建“人机协同”（Human-in-the-loop）的下一代科学智能体（AI Agents），推动数据驱动的自主科学发现。
核心研究方向：
1. 单分子分析的深度学习基础模型 针对单分子数据信噪比低、异质性强等分析挑战，课题组开发了一系列先进的深度学习框架（如 META-SiM、Kin-SiM 和 AutoSiM）。通过引入基于Transformer架构的基础模型和多任务预训练策略，我们实现了单分子轨迹分类、分割、状态理想化及动力学参数提取的全面自动化。这些工具在准确率对齐人类专家的同时，不仅将分析效率提升了百倍以上，更成功助力发现了以往被掩盖的稀有生物物理现象（如未知的RNA剪接中间态）。
2. 单分子生物物理与核酸纳米技术 利用高时空分辨的单分子技术（如smFRET和单粒子追踪），我们致力于解析纳米尺度下的复杂分子动态互作。代表性工作包括探究DNA分子机器的动力学运动机理，以及开发优化基于单分子动力学指纹（kinetic fingerprint）的高特异性、免扩增核酸检测技术，为重大疾病的早期分子诊断与新型纳米器件设计提供重要基础。
3. 面向科学发现的下一代 AI 智能体 为打破单分子数据分析的专业壁垒，课题组正前瞻性地构建大语言模型驱动的双智能体（Dual-Agent）架构。系统协同负责基础模型训练与微调的“机器学习（ML）智能体”，以及负责原始实验数据解析的“数据科学（DS）智能体”。该架构能够针对多样的单分子实验系统，自动生成定制化的高性能数据分析工具，极大加速微观生物分子动力学规律的挖掘.
Single-molecule microscopy (e.g., Single-Molecule Fluorescence Microscopy, SMFM) has emerged as a transformative tool in unraveling intricate molecular processes. By removing the limitations imposed by ensemble averaging, it allows for the direct observation of transient intermediates and rare biomolecular dynamics at the nanoscale and millisecond time scales that would otherwise be masked. Our research group is dedicated to synergizing cutting-edge Artificial Intelligence (AI) and single-molecule technologies. We aim to develop generalizable, scalable, and automated "meta-tools" to overcome the poor adaptability, high labor costs, and user bias inherent in traditional single-molecule data analysis. By doing so, we strive to accurately unveil transient dynamics in complex biological and nanotechnology systems. Furthermore, we are pioneering human-in-the-loop AI agents to fully automate data-driven scientific discoveries.
Core Research Areas:
1. Deep Learning Foundation Models for Single-Molecule Analysis To address the challenges of low signal-to-noise ratios and high heterogeneity in single-molecule data, we have developed a suite of advanced deep learning frameworks (e.g., META-SiM, Kin-SiM, and AutoSiM). By leveraging Transformer-based foundation models and multi-task pretraining, we achieve fully automated trace classification, segmentation, idealization, and kinetic extraction. These tools operate orders of magnitude faster than manual analysis while maintaining expert-level accuracy, successfully facilitating the discovery of rare, previously undetected biomolecular states, such as novel pre-mRNA splicing intermediates.
2. Single-Molecule Biophysics and Nucleic Acid Nanotechnology Utilizing high spatiotemporal resolution techniques like smFRET and single-particle tracking, we explore complex molecular interactions at the nanoscale. Representative work includes probing the kinetic speed limits of DNA molecular machines and optimizing Single-Molecule Recognition through Equilibrium Poisson Sampling (SiMREPS) for highly specific, amplification-free nucleic acid biosensing, providing new strategies for early molecular diagnostics.
3. Next-Generation AI Agents for Scientific Discovery To democratize advanced single-molecule analytics, we are developing a forward-looking, LLM-powered Dual-Agent architecture. By orchestrating a "Machine Learning Agent" for foundation model training and a "Data Science Agent" for raw data interpretation, this system automatically generates customized analytical tools tailored to diverse experimental setups, significantly accelerating the extraction of scientific insights.

代表性论文(^# co-first author, ^* Corresponding author):

1. Li, J. ^#*, Zhang, L.^#, Johnson-Buck, A., & Walter, N. G.^* (2025). Foundation model for efficient biological discovery in single-molecule time traces. Nature Methods, 22(10), 2149-2160.
2. Zhang, L^#., Li, J^#*. and Walter, N.G.^*, (2025). Pretrained deep neural network Kin-SiM for single-molecule FRET trace idealization. The Journal of Physical Chemistry B, 129 (4), 1167-1175.
3. Song, D.; Zhang, X.; Li, B.; Sun, Y.; Mei, H.; Cheng, X.; Li, J.^*; Cheng, X.^*; Fang, N.^* (2024) Deep learning-assisted automated multidimensional single particle tracking in living cells. Nano Letters, 24 (10), 3082-3088.
4. Yang, M.; Mansour, N.; Huang, T.; Heffer, D.; Pei, Y.; Huang, W.; Li, J.^*; Dong, B.^*; Fang, N.^* (2024). Reaction Initiated Single Molecule Tracking of Mass Transfer in Core-Shell Mesoporous Silica Particle. Analytical Chemistry, 96 (6), 2500-2505.
5. Huang, Y., Hu, Y., Yuill, E. M., Marriott, A. S., Chadwick, J., Li, J., ... & Miller, S. A. (2022). Circumventing glass vial and diluent effects on solution stability of small molecule analytes during analytical method development and validation. Journal of Pharmaceutical and Biomedical Analysis, 213, 114676.
6. Yuill, E. M., Ileka, K. M., La Cruz, T. E., Li, J., Shackman, J. G., Tattersall, P. I., & Zang, J. (2021). Leveraging AQbD Principles for Development of Challenging Drug Substance Stability-Indicating Methods. Organic Process Research & Development, 25(6), 1431-1439.
7. Li, J. ^#, Zhang, L. ^#, Johnson-Buck, A., & Walter, N. G. (2020). Automatic classification and segmentation of single-molecule fluorescence time traces with deep learning. Nature Communications, 11(1), 1-11.
    ○National Cancer Institute’s Epidemiology and Genomics Research Program (EGRP) Research Highlights for 2020 (link).
8. Johnson-Buck, A. ^#, Li, J. ^#, Tewari, M., & Walter, N. G. (2019) A guide to nucleic acid detection through single molecule kinetic fingerprinting”, Methods 153, 3-12
9. Li, J. ^#, Johnson-Buck, A. ^#, Yang, Y. R., Shih, W. M., Yan, H., & Walter, N. G. (2018) Exploring the speed limit of toehold exchange with a cartwheeling DNA acrobat. Nature Nanotechnology 13(8), 723–729.
    ○Highlighted on Nature: “Gymnastic feats help DNA ‘walker’ set speed record” DOI: 10.1038/d41586-018-05127-8
    ○Highlighted on Nature Reviews Materials: “Head over heels”. 3, 155 (2018).
10. Lin, H. ^#, Li, J. ^#, Liu, B., Liu, D., Liu, J., Terfort, A., ... & Ren, B. (2013). Uniform gold spherical particles for single-particle surface-enhanced Raman spectroscopy Physical Chemistry Chemical Physics, 15(12), 4130-4135

荣誉、奖励及参加学术团体的情况:

2016 Chemistry Department Fellowship, University of Michigan
2016 Student Travel Award funded by IUPAB for the Biophysical Society Meeting, Biophysical Society
2015 Future Faculty Graduate Student Instructor Fellowship, University of Michigan
2011 China National Endeavor Scholarship, Xiamen University