Carlwanghy1128/DGAT

Introduction

Spatial transcriptomics (ST) technologies provide genome-wide mRNA profiles in tissue context but lack direct protein-level measurements, which are critical for interpreting cellular function and microenvironmental organization. We present DGAT (Dual-Graph Attention Network), a deep learning framework that imputes spatial protein expression from transcriptomics-only ST data by learning RNA–protein relationships from spatial CITE-seq datasets. DGAT constructs heterogeneous graphs integrating transcriptomic, proteomic, and spatial information, encoded using graph attention networks. Task-specific decoders reconstruct mRNA and predict protein abundance from a shared latent representation. Benchmarking across public and in-house datasets—including tonsil, breast cancer, glioblastoma, and malignant mesothelioma— demonstrates that DGAT outperforms existing methods in protein imputation accuracy. Applied to ST datasets lacking protein measurements, DGAT reveals spatially distinct cell states, immune phenotypes, and tissue architectures not evident from transcriptomics alone. DGAT enables proteome-level insights from transcriptomics-only data, bridging a critical gap in spatial omics and enhancing functional interpretation in cancer, immunology, and precision medicine.

Data

The training datasets for DGAT model, pre-trained model and Spatial Transcriptomics for predictions can be downloaded here.

For prediction using pretrainde model, please download DGAT_prediction_ST_data, DGAT_pretrained_models to the same path as this Readme.

For training, please download DGAT_training_datasets to the same path as this Readme.

For reproduction, please download DGAT_results_reproduction to ./Reproduction/results/ before you run the scripts and notebooks.

Installation

The code runs on Python 3.11.

You can install the required packages using pip:

• For CUDA

  pip install -r requirements_CUDA.txt
  

• For CPU-only

  pip install -r requirements_CPU.txt
  

• For already installed torch

  pip install -r requirements_torch_ready.txt
  

10-minute Quick Start

Train

Demo1_Train notebook will lead you through the training process of DGAT model on one sample. Detailed instructions are provided in the notebook.

Predict

To run the prediction demo after Demo1_Train notebook, you can follow the Demo2_Predict notebook. This notebook will guide you through the process of using the pre-trained DGAT model from Demo1_Train notebook to predict gene expression from Spatial Transcriptomics data and the downstream analysis.

Further Explore

For further exploring, such as training on multiple samples from our datasets or yours, please follow the Pretrain_DGAT notebook.

Demo3_Predict_ST notebook will guide you through the process of predicting protein expressions on ST datasets using the pre-trained DGAT model.

Citation

@article{wang2025dgat,
title={{DGAT}: A Dual-Graph Attention Network for Inferring Spatial Protein Landscapes from Transcriptomics},
author={Wang, Haoyu and Cody, Brittany and Osmanbeyoglu, Hatice Ulku},
journal={bioRxiv},
year={2025},
doi={10.1101/2025.07.05.662121},
note = {preprint}
}