Xiaohan Kuang

Vanderbilt University - Data Science Institute

I am a graduate student at Vanderbilt University's Data Science Institute specializing in machine learning and its applications in computational biology. I completed my Bachelor of Science in Computer Science, Mathematics, and Data Science from the University of Wisconsin–Madison in 2022.

Research Interest
Integrating advanced data science methodologies and AI-driven models, such as diffusion models and equivariant graph neural networks, with biological research and other interdisciplinary fields to improve current methods and develop practical applications.

Experience

Vanderbilt University Medical Center, Meiler Lab

Research Assistant
  • Developed a novel generative AI framework that integrates score-based diffusion models with equivariant graph neural networks to predict water or metal-binding sites in protein structures, outperforming state-of-the-art tools. Two manuscripts have resulted from this project: one accepted at NeurIPS MLSB 2024, and the other currently under review.
  • Developed a computational approach to detect antigen-binding sites by representing antigen surface information as 2D images. Preliminary results indicate that convolutional neural networks (CNNs) can predict binding sites effectively, achieving an AUC greater than 0.93. Part of this work has been published in a peer-reviewed journal.
  • Developed an active learning framework integrating molecular dynamics simulations and machine learning to explore the relationship between peptide sequences and their fibril formation propensity. This manuscript is currently under review.
Oct. 2023 - Present

Vanderbilt University, Data Science Institute

Researcher
  • Developed an AI-powered web application to convert complex academic papers into easily understandable summaries, bridging the gap between scientific research and the general public.
  • Implemented the platform using Streamlit, integrating user authentication and content customization features with Python and Deta to provide a seamless and engaging user experience.
Fall 2023

Publications

  1. Kuang, X., Su, Z., Liu, Y. (Lance), Lin, X., Spencer-Smith, J., Derr, T., Wu, Y., & Meiler, J. SuperWater: Predicting water molecule positions on protein structures by generative AI. Preprint: bioRxiv (2024)
  2. Lin, X., Su, Z., Liu, Y., Liu, J., Kuang, X., Cummings, P. T., Spencer-Smith, J., & Meiler, J. SuperMetal: A generative AI framework for rapid and precise metal ion location prediction in proteins. Accepted at NeurIPS MLSB (2024). Poster: MLSB
  3. Kuang, X., Jalali, S., Rahman, T., Michalowski, J., Sheng-Wong, C., Su, Z., & Dias, C. L. Discovering new amyloid-like peptides using all-atom simulations and artificial intelligence. Preprint: bioRxiv (2024)
  4. Zhang, G., Kuang, X., Zhang, Y., Liu, Y., Su, Z., Zhang, T., & Wu, Y. Machine-learning-based structural analysis of interactions between antibodies and antigens. Published in BioSystem (2024) [Paper]

Interests

When I’m not immersed in research or coding, I enjoy exploring hobbies that keep me balanced and inspired. Practicing Chinese calligraphy for over 14 years has been a journey of creativity and mindfulness, offering a connection to tradition through brushstroke. Music is another passion - I often pick up my guitar to teach myself new songs, discovering melodies one chord at a time. My interests constantly evolve, and I’m always curious to try something new that sparks joy and creativity.