Bohan Li

I am a Ph.D. candidate at Beihang University, working on generative modeling for spatial omics.

My research focuses on resolution enhancement and continuous modeling of biological systems, aiming to reconstruct high-resolution molecular landscapes from sparse and partially observed data.

Research Focus

  • Generative modeling for spatial omics
  • Continuous representations of biological systems
  • Multi-modal integration under partial observations

Selected Publications

Nature Communications (2024)
Tissue characterization at an enhanced resolution across spatial omics platforms with deep generative model
https://doi.org/10.1038/s41467-024-50837-5

Cell (2026)
Whole-body molecular and cellular mapping of the laboratory mouse
https://doi.org/10.1016/j.cell.2026.03.006

Under review at Nature Methods
Reconstructing isotropic-resolution spatial transcriptomics from serial sections by modeling tissue continuity
https://doi.org/10.1101/2025.08.15.670472

→ Full publication list

Research Projects

soScope — Lateral resolution enhancement

A generative framework modeling spatial omics measurements as aggregated observations of latent cellular states.

  • Variational inference formulation of spatial aggregation
  • Unified modeling of histology and gene expression
  • Scalable reconstruction pipeline

GitHub

isoST — Depth-wise resolution enhancement

A continuous generative framework based on stochastic differential equations to model gene expression along tissue depth.

  • Continuous modeling of biological structures
  • Reconstruction from sparse serial sections
  • Isotropic-resolution inference

GitHub

LABEL — Cross-scale annotation transfer

Framework for transferring molecular annotations to histological images.

  • Cross-scale alignment between omics and imaging
  • Model design and system integration
  • Large-scale dataset validation

GitHub

Academic Service

  • Invited reviewer for Briefings in Bioinformatics