← All research themes
Spatial transcriptomics

Spatial transcriptomics

We predict spatial gene expression from pathology images and study cell types and differential expression in tissue.

Selected Papers

Selected papers

Selected NeurIPS 2026 · Differential gene rankings

Preserving DEG Rankings for Gene Discovery in Histology-Based Spatial Gene Expression Prediction

This study focuses on preserving differential gene expression rankings for gene discovery from histology.

Cell-type prototypes and pathology image-based gene expression estimation
Fig. 2 · Cell-type prototypes for gene expression prediction · Source: paper PDF ↗
Selected CVPR 2026 · Image-to-expression prediction

Cell-Type Prototype-Informed Neural Network for Gene Expression Estimation from Pathology Images

Cell-type gene expression prototypes provide an interpretable intermediate representation for predicting gene expression from pathology images.

Paper and abstract ↗
Fig. 1 · Concept of auxiliary gene learning
Fig. 1 · Concept of auxiliary gene learning · Source: paper PDF ↗
Selected AAAI 2026 · Auxiliary gene selection

Auxiliary Gene Learning: Spatial Gene Expression Estimation by Auxiliary Gene Selection

Auxiliary genes outside the target set are selected as useful training tasks to improve spatial expression prediction.

Paper and abstract ↗
Fig. 1 · Relative expression under batch effects and noise
Fig. 1 · Relative expression under batch effects and noise · Source: paper PDF ↗
Selected NeurIPS 2025 · Relative expression trends

Learning Relative Gene Expression Trends from Pathology Images in Spatial Transcriptomics

Paper and abstract ↗

The model learns relative gene expression trends despite batch effects and measurement noise.

Publications

Related publications

  1. Preserving DEG Rankings for Gene Discovery in Histology-Based Spatial Gene Expression Prediction
    Kaito Shiku, Kazuya Nishimura, Yasuhiro Kojima, and Ryoma Bise · The Fortieth Annual Conference on Neural Information Processing Systems (NeurIPS), 2026.
  2. Leveraging Cell-level Spatial Transcriptomics as Weak Supervision for Nuclei Classification
    Kazuya Nishimura, Ryoma Bise, Haruka Hirose, and Yasuhiro Kojima · International Workshop on Medical Imaging Analysis for Spatial Omics (MISO), MICCAI 2026 Workshop.
  3. Cell-Type Prototype-Informed Neural Network for Gene Expression Estimation from Pathology Images
    Kazuya Nishimura, Ryoma Bise, Shinnosuke Matsuo, Haruka Hirose, and Yasuhiro Kojima · The IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), 2026.
  4. Auxiliary Gene Learning: Spatial Gene Expression Estimation by Auxiliary Gene Selection
    Kaito Shiku, Kazuya Nishimura, Shinnosuke Matsuo, Yasuhiro Kojima, and Ryoma Bise · The 40th Annual AAAI Conference on Artificial Intelligence (AAAI), 2026.
  5. Learning to Relative Expression under Batch Effects and Stochastic Noise in Spatial Transcriptomics
    Kazuya Nishimura, Haruka Hirose, Ryoma Bise, Kaito Shiku, Yasuhiro Kojima · The Thirty-Ninth Annual Conference on Neural Information Processing Systems (NeurIPS), 2025.
  6. Towards Spatial Transcriptomics-Guided Pathological Image Recognition With Batch-Agnostic Encoder
    Kazuya Nishimura, Ryoma Bise, and Yasuhiro Kojima · IEEE International Symposium on Biomedical Imaging (ISBI), 2025.

View all publications →