If you find our work useful in your research, please consider citing our SpatialCTD paper:
@article{ding2024spatialctd,
title={SpatialCTD: A Large-Scale Tumor Microenvironment Spatial Transcriptomic Dataset to Evaluate Cell Type Deconvolution for Immuno-Oncology},
author={Ding, Jiayuan and Li, Lingxiao and Lu, Qiaolin and Venegas, Julian and Wang, Yixin and Wu, Lidan and Jin, Wei and Wen, Hongzhi and Liu, Renming and Tang, Wenzhuo and others},
journal={Journal of Computational Biology},
year={2024},
publisher={Mary Ann Liebert, Inc., publishers 140 Huguenot Street, 3rd Floor New~…}
}
We have made jupiter notebooks with detailed instructions for GNNDeconvolver running. Please refer to the jupiter notebooks under GNNDeconvolver folder.
SpatialCTD is a large-scale spatial transcriptomic dataset encompassing 1.8 million cells from the human tumor microenvironment across the lung, kidney, and liver. The ST datasets used for CTD in SpatialCTD v1 are generated from real spatial transcriptomic datasets via gridded spots method.
For each sample in SpatialCTD, it includes:
The dataset generation workflow and summary are as follows:
For thorough details, see the preprint: [Biorxiv]
To ease the process of converting spatial transcriptomics data at single-cell resolution to pseudo spots for benchmarking cell type deconvolution, we have developed a web-based online conversion tool. The online converter tool can be accessed at [website]
A few required inputs are as below: