Skip to content

This tool receives a normalized h5ad matrix of myeloid cells and calculates the usage of the consensus myeloid cNMF programs in gliomas by the cells in the uploaded h5ad matrix.

Notifications You must be signed in to change notification settings

cwru-non-academic/Calculate_Myeloid_cNMF_Usage

 
 

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

40 Commits
 
 
 
 
 
 

Repository files navigation

Calculate_Myeloid_cNMF_Usage

This shiny app receives a gene expression matrix and calculates the usage of the cell annotation cNMF programs, and the consensus myeloid cNMF programs in gliomas.

The online version of this tool is available at: [(https://consensus-myeloid-program-calculator.shinyapps.io/shinyapp/)]

However, it is recommended that you download this app to avoid server delays and if you have large matrices.

The use of this tool is simple: Make sure to install all the required dependencies and follow the instructions below.

Dependencies:

R > 4.0 (The latest version of RStudio is highly recommended)

Python > 3.7

The following R Libraries:

shiny

shinyFiles

reticulate

App Setup:

1- Install the required R libraries (Needs to be done one time in each system):

install.packages("shiny")

install.packages("shinyFiles")

install.packages("reticulate")

reticulate::install_python(version = '3.9')

2- Download the four required files ("app.R", "Myeloid_NMF_Average_Gene_Spectra.txt", "cnmf_run.spectra.k_18.dt_0_015.consensus.df.npz", and "your_python_script_module.py") (Located in the "Required Files" folder) and place them together in one folder (This folder should be your working directory in R when you want to use the app). Ensure you download the files correctly through Git Hub (DO NOT right-click on the file to download it).

3- Start your R session and set the working directory to the folder mentioned above in step 2 using setwd().

4- Load the required libraries as follows:

library(shiny)

library(shinyFiles)

5- Then type the following in R:

shiny::runApp()

6- An interactive toolbox will be generated if everything is set up well.

7- Choose Annotation mode or Myeloid Calculation mode.

8— Upload the h5ad, csv or mtx matrix by clicking on browse and waiting for the "Calculating usage" bar to complete. (It should take a few minutes, depending on the size of the matrix.). If you have chosen mtx format, you will need to upload the features and barcodes files which are additional outputs of CellRanger or STARSolo.

9- Click "Download Output Data" to download the calculated usage.

Input:

1- Choose the mode. "Annotation Mode" calculates the enrichment of cell types NMF programs to help you annotate the Cells. Myeloid Program Calculation Mode calculates the usages of the consensus cNMF programs in glioma-associated myeloid cells. Upload gene expression matrix of myeloid cells for Myeloid Program Calculation Mode. a. Choose the input format after choosing the mode. The matrix should be in mtx, h5ad, or csv formats. For CSV, genes should be in rows and cells in columns. For mtx format, barcodes and feature files have to be uploaded as well. They should be in the format accepted by Scanpy. Feature file has to have three columns, with the second column including gene symbols and the third column having the words Gene Expression in all rows b. The matrix can be normalized or raw. c. mtx and associated files can be gzipped or uncompressed. Uploading a gzipped mtx file is highly recommended as it saves time.

2- Scanpy generated h5ad is highly recommended.

3- The values in adata.X can be normalized (It is recommended to use sc.pp.normalize_per_cell) or raw.

4- Genes should be stored as official gene symbols (HGNC official gene symbols).

Details are provided in the interactive toolbox.

Output:

1-The output is a data frame in which the rows denote cells existing in the input h5ad matrix, whereas the columns represent the indicated consensus myeloid cNMF programs in human gliomas.

2- IT IS IMPORTANT TO NOTE THAT THE USAGE VALUES IN THE OUTPUT ARE NORMALIZED AS PERCENTAGES PER CELL TO ENABLE COMPARISON AMONG CELLS

Details are provided in the interactive toolbox.

Demo Data:

Instructions for using the demo data are included in the "Demo Data" folder.

Contact:

This shiny app is developed and maintained by Chadi A. El Farran, M.Sc., Ph.D. ([email protected]), with significant contributions from Charles P. Couturier, MD, Ph.D., and Tyler E. Miller, MD, Ph.D.

If you use this tool, please cite:

About

This tool receives a normalized h5ad matrix of myeloid cells and calculates the usage of the consensus myeloid cNMF programs in gliomas by the cells in the uploaded h5ad matrix.

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages

  • Python 51.3%
  • R 48.7%