scDeepSort

Reference-free Cell-type Annotation for Single-cell Transcriptomics using Deep Learning with a Weighted Graph Neural Network

Recent advance in single-cell RNA sequencing (scRNA-seq) has enabled large-scale transcriptional characterization of thousands of cells in multiple complex tissues, in which accurate cell type identification becomes the prerequisite and vital step for scRNA-seq studies.

To addresses this challenge, we developed a reference-free cell-type annotation method, namely scDeepSort, using a state-of-the-art deep learning algorithm, i.e. a modified graph neural network (GNN) model. It’s the first time that GNN is introduced into scRNA-seq studies and demonstrate its ground-breaking performances in this application scenario. In brief, scDeepSort was constructed based on our weighted GNN framework and was then learned in two embedded high-quality scRNA-seq atlases containing 764,741 cells across 88 tissues of human and mouse, which are the most comprehensive multiple-organs scRNA-seq data resources to date. For more information, please refer to a preprint in bioRxiv 2020.05.13.094953.

Install

1. Download source codes of scDeepSort.
2. Download pretrained models from the release page and uncompress them.

tar -xzvf pretrained.tar.gz

After executing the above steps, the final scDeepSort tree should look like this:

 |- pretrianed
     |- human
        |- graphs
        |- statistics
        |- models
     |- mouse
        |- graphs
        |- statistics
        |- models
 |- test
     |- human
     |- mouse
 |- map
    |- human
        |- map.xlsx
    |- mouse
        |- map.xlsx
 |- models
    |- __init__.py
    |- gnn.py
 |- utils
    |- __init__.py
    |- preprocess.py
 |- run.py
 |- requirements.txt
 |- README.md

Dependency

• Dependencies can also be installed using pip install -r requirements.txt

Usage

Prepare test data

1. The file name of test data should be named in this format: species_TissueNumber_data.csv. For example, human_Pancreas11_data.csv is a data file containing 11 human pancreas cells.

2. The test single-cell transcriptomics csv data file should be normalized with the defalut LogNormalize method with Seurat (R package), wherein the column represents each cell and the row represent each gene, as shown below.

   	Cell 1	Cell 2	Cell 3	...
   Gene 1	0	2.4	5.0	...
   Gene 2	0.8	1.1	4.3	...
   Gene 3	1.8	0	0	...
   ...	...	...	...	...

3. All the test data should be included under the test directory. Furthermore, all of the human testing datasets and mouse testing datasets are required to be under ./test/human and ./test/mouse respectively.

Run

Evaluate

To evaluate one data file mouse_Liver4122_data.csv, which we report the prediction accuracy in the paper, you should execute the following command:

python run.py --species human --tissue Liver --test_dataset 4122 --gpu -1 --evaluate

• --species The species of cells, human or mouse.
• --tissue The tissue of cells. see Details
• --test_dataset The dataset to be tested, in other words, as the file naming rule states, it is exactly the number of cells in the data file.
• --gpu Specify the GPU to use, -1 for cpu.

Test

To test one data file human_Pancreas11.csv, you should execute the following command:

python run.py --species human --tissue Pancreas --test_dataset 11 --gpu -1 --test

• --species The species of cells, human or mouse.
• --tissue The tissue of cells. see Details
• --test_dataset The dataset to be tested, in other words, as the file naming rule states, it is exactly the number of cells in the data file.
• --gpu Specify the GPU to use, -1 for cpu.

Output

For each dataset, it will output a .csv file named as species_Tissue_Number.csv under the result directory. For example, output of test dataset human_Pancreas11_data.csv is human_Pancreas_11.csv

Evaluate:

Each line of the output file corresponds to the original cell type and predictive cell type.

Test:

Each line of the output file corresponds to the predictive cell type.

Details

• Adipose
• Adrenal_gland
• Artery
• Ascending_colon
• Bladder
• Blood
• Bone_marrow
• Brain
• Cervix
• Chorionic_villus
• Colorectum
• Cord_blood
• Epityphlon
• Esophagus
• Fallopian_tube
• Female_gonad
• Fetal_adrenal_gland
• Fetal_brain
• Fetal_calvaria
• Fetal_eye
• Fetal_heart
• Fetal_intestine
• Fetal_kidney
• Fetal_liver
• Fetal_Lung
• Fetal_male_gonad
• Fetal_muscle
• Fetal_pancreas
• Female_gonad
• Fetal_rib
• Fetal_skin
• Fetal_spinal_cord
• Fetal_stomach
• Fetal_thymus
• Gall_bladder
• Heart
• Kidney
• Liver
• Lung
• Muscle
• Neonatal_adrenal_gland
• Omentum
• Pancreas
• Placenta
• Pleura
• Prostat
• Spleen
• Stomach
• Temporal_lobe
• Thyroid
• Trachea
• Ureter

• Bladder
• Blood
• Bone_marrow
• Bone_Marrow_mesenchyme
• Brain
• Embryonic_mesenchyme
• Fetal_brain
• Fetal_intestine
• Fetal_liver
• Fetal_lung
• Fetal_stomach
• Intestine
• Kidney
• Liver
• Lung
• Mammary_gland
• Muscle
• Neonatal_calvaria
• Neonatal_heart
• Neonatal_muscle
• Neonatal_pancreas
• Neonatal_rib
• Neonatal_skin
• Ovary
• Pancreas
• Placenta
• Prostate
• Spleen
• Stomach
• Testis
• Thymus
• Uterus

Examples

python run.py --species human --tissue Pancreas --test_dataset 11 --gpu -1 --test

python run.py --species mouse --tissue Intestine --test_dataset 28 --gpu -1 --test

To test all datasets, please download human_test_data and mouse_test_data in release page