Benchmark-CTCM-ST

Datasets for our new paper: "Benchmarking cell-type clustering methods for spatially resolved transcriptomics data" https://doi.org/10.1093/bib/bbac475.

If you find this repo useful please cite:

@article{10.1093/bib/bbac475,
    author = {Cheng, Andrew and Hu, Guanyu and Li, Wei Vivian},
    title = "{Benchmarking cell-type clustering methods for spatially resolved transcriptomics data}",
    journal = {Briefings in Bioinformatics},
    year = {2022},
    month = {11},
    issn = {1477-4054},
    doi = {10.1093/bib/bbac475}
}

0. Sample folder structure

Benchmark-CTCM-ST
│   README.md
└───Dataset1/
│   └───images_sd0.5/
│       │  rgb_img1.png #Simulated images for replicate 1 with standard deviation 0.5
│       │  rgb_img2.png
│       │   ...
│   └───images_sd5/
│       │  rgb_img1.png
│       │   ...
│   └───images_sd10/
│       │  rgb_img1.png
│       │   ...
│   └───images_sd25/
│       │  rgb_img1.png
│       │   ...
│   └───images_sd50/
│       │  rgb_img1.png
│       │   ...
│   └───simcounts/
│       │   Dataset1_counts1.rds #Simulated counts for replicate 1
│       │   Dataset1_counts2.rds
│       │   ...
│   └───spatial_info/
│       │   spatial_trans1.tsv #Pixel-mapped spatial coordinates
│       │   spatial_trans2.tsv
│       │   ...
│   └───true_cl/
│       │   Dataset1_true_cl1.rds #Cell types assigned through RShiny
│       │   Dataset1_true_cl2.rds
│       │   ...

1. Simulated counts

A .rds file containing:

• Genes X Cell matrix (dgCMatrix)

  2. Spatial location information

  A .tsv file containing 3 columns:

• cell index
• X0: x-coordinate pixel-mapped values
• X1: y-coordinate pixel-mapped values

3. Simulated histology images

A .png image with shape (288, 288, 3) generated with a standard deviations of 0.5, 5, 10, 25, and 50.

4. true_cl: Ground truth cell types assigned through RShiny

• Factor with k levels, where k is the number of unique cell types assigned.