# CellCover
This is the python version of CellCover. To run CellCover, Gurobi installation is necessary. Please follow the README.md in [lanlanji/CoveringPackage (github.com)](https://github.com/lanlanji/CoveringPackage) to acquire and install the Gurobi academic liscence.
### Installing CellCover:
```python
pip install CellCover
```
### Using CellCover to obtain marker panel
To run the CelCover, the following python variables need to be define first:
- **data**: a numpy array of your single cell RNA-seq data with shape (N,G) where N is the number of cells and G is the size of the gene portfolio.
- **gene**: a numpy array of gene names with shape (G,)
- **CellTypeLabels**: a numpy array of cell types with shape (N,)
- **CellTypeNames**: a numpy array of distinct cell type names (string)
- **ct**: a single string of the cell type name that user want to find covering markers for, e.g. "CD4"
What is more, there is a list of hyperparameters that need to be defined before running the covering:
- **binarization_threshold**: the threshold above which we binarize the gene expression to 1, below which we binarize the gene expression to 0
- **minSize**: the depth of covering
- **alpha**: 1 - covering rate. The default is $0.05$
- **te**: This is a parameter for pruning the data. For each cell type, the gene expressing more than te * 100 percent of time are selected for finding the covering markers. The default is $0.1$.
- **top_num_gene**: This is another pruning parameter. In each class, **top_num_gene** number of genes with the highest margin score will be selected for marker selection. The default is $6000$.
The pipeline of getting the covering marker panel of the user defined cell type **ct** is
```python
from CellCover import binarization
from CellCover import SensList
from CellCover import weight
from CellCover import covering
from CellCover import getCoveringVariables
data = binarization(mat = data, binarization_threshold = binarization_threshold)
sens = SensList(mat = data, CellTypeLabels = CellTypeLabels, CellTypeNames=CellTypeNames)
X,w,g = weight(mat = data,sens =sens, CellTypeLabels = CellTypeLabels, CellTypeNames = CellTypeNames, ct = ct,gene = gene)
cov = covering(Z=X, minSize = minSize, alpha = alpha,weights = w)
marker = getCoveringVariables(cov, ngenes = len(g), geneNames = g, nlevels = 1)
```
Raw data
{
"_id": null,
"home_page": "",
"name": "CellCover",
"maintainer": "",
"docs_url": null,
"requires_python": "",
"maintainer_email": "",
"keywords": "python,single-cell,marker gene",
"author": "Laurent Younes",
"author_email": "<laurent.younes@jhu.edu>",
"download_url": "https://files.pythonhosted.org/packages/ef/88/f6298dd86ae90fbe8e04a0e05a9e4bc24fa9139e7968638156280a76ea8b/CellCover-0.1.9.tar.gz",
"platform": null,
"description": "\n# CellCover\n\nThis is the python version of CellCover. To run CellCover, Gurobi installation is necessary. Please follow the README.md in [lanlanji/CoveringPackage (github.com)](https://github.com/lanlanji/CoveringPackage) to acquire and install the Gurobi academic liscence.\n\n### Installing CellCover:\n\n```python\npip install CellCover\n```\n\n### Using CellCover to obtain marker panel\n\nTo run the CelCover, the following python variables need to be define first: \n\n- **data**: a numpy array of your single cell RNA-seq data with shape (N,G) where N is the number of cells and G is the size of the gene portfolio.\n\n- **gene**: a numpy array of gene names with shape (G,)\n\n- **CellTypeLabels**: a numpy array of cell types with shape (N,)\n\n- **CellTypeNames**: a numpy array of distinct cell type names (string)\n\n- **ct**: a single string of the cell type name that user want to find covering markers for, e.g. \"CD4\"\n\nWhat is more, there is a list of hyperparameters that need to be defined before running the covering:\n\n- **binarization_threshold**: the threshold above which we binarize the gene expression to 1, below which we binarize the gene expression to 0 \n\n- **minSize**: the depth of covering\n\n- **alpha**: 1 - covering rate. The default is $0.05$\n\n- **te**: This is a parameter for pruning the data. For each cell type, the gene expressing more than\u00a0te\u00a0* 100 percent of time are selected for finding the covering markers. The default is\u00a0$0.1$.\n\n- **top_num_gene**: This is another pruning parameter. In each class,\u00a0**top_num_gene**\u00a0number of genes with the highest margin score will be selected for marker selection. The default is $6000$.\n\nThe pipeline of getting the covering marker panel of the user defined cell type **ct** is\n\n```python\nfrom CellCover import binarization\nfrom CellCover import SensList\nfrom CellCover import weight\nfrom CellCover import covering\nfrom CellCover import getCoveringVariables\ndata = binarization(mat = data, binarization_threshold = binarization_threshold)\nsens = SensList(mat = data, CellTypeLabels = CellTypeLabels, CellTypeNames=CellTypeNames)\nX,w,g = weight(mat = data,sens =sens, CellTypeLabels = CellTypeLabels, CellTypeNames = CellTypeNames, ct = ct,gene = gene)\ncov = covering(Z=X, minSize = minSize, alpha = alpha,weights = w)\nmarker = getCoveringVariables(cov, ngenes = len(g), geneNames = g, nlevels = 1)\n\n```\n\n",
"bugtrack_url": null,
"license": "",
"summary": "CellCover",
"version": "0.1.9",
"project_urls": null,
"split_keywords": [
"python",
"single-cell",
"marker gene"
],
"urls": [
{
"comment_text": "",
"digests": {
"blake2b_256": "414af9bd6e0fb30f3c269574504202f3a910e48ddf8c2e274fa80b430b51b88c",
"md5": "5ae4cc21f25bccbce72024f477fd629d",
"sha256": "c8e504c75815e6e68e7ce73cb5c6ef13420fdffc13e08e18f7fd3c2283674f12"
},
"downloads": -1,
"filename": "CellCover-0.1.9-py2.py3-none-any.whl",
"has_sig": false,
"md5_digest": "5ae4cc21f25bccbce72024f477fd629d",
"packagetype": "bdist_wheel",
"python_version": "py2.py3",
"requires_python": null,
"size": 4145,
"upload_time": "2023-08-17T20:30:32",
"upload_time_iso_8601": "2023-08-17T20:30:32.851036Z",
"url": "https://files.pythonhosted.org/packages/41/4a/f9bd6e0fb30f3c269574504202f3a910e48ddf8c2e274fa80b430b51b88c/CellCover-0.1.9-py2.py3-none-any.whl",
"yanked": false,
"yanked_reason": null
},
{
"comment_text": "",
"digests": {
"blake2b_256": "ef88f6298dd86ae90fbe8e04a0e05a9e4bc24fa9139e7968638156280a76ea8b",
"md5": "9bd8b87e09f8c67f4c5eb0f4ec7e74da",
"sha256": "68310a770ef3509340767f459674a043ccbcaec8b1f1b791799d8af3fe6c7a6f"
},
"downloads": -1,
"filename": "CellCover-0.1.9.tar.gz",
"has_sig": false,
"md5_digest": "9bd8b87e09f8c67f4c5eb0f4ec7e74da",
"packagetype": "sdist",
"python_version": "source",
"requires_python": null,
"size": 3929,
"upload_time": "2023-08-17T20:30:34",
"upload_time_iso_8601": "2023-08-17T20:30:34.604525Z",
"url": "https://files.pythonhosted.org/packages/ef/88/f6298dd86ae90fbe8e04a0e05a9e4bc24fa9139e7968638156280a76ea8b/CellCover-0.1.9.tar.gz",
"yanked": false,
"yanked_reason": null
}
],
"upload_time": "2023-08-17 20:30:34",
"github": false,
"gitlab": false,
"bitbucket": false,
"codeberg": false,
"lcname": "cellcover"
}
JSON