Skip to content

Predicting missing cell-type proportions from deconvolution residual using NNLS.

License

Notifications You must be signed in to change notification settings

greenelab/pred_missing_celltypes

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

39 Commits
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

pred_missing_celltypes

Predicting missing cell-types from deconvolution reference using residual.

This repository is currently a work in progress of a project aimed at uncovering information from missing cell types, after RNA-bulks have been deconvoluted. Our study aims to answer the question of what happens when we have real and accumulated cells missing from the single-cell dataset in bulk deconvolution methodologies, and whether we are able to retrieve the missing cell-type’s proportions through the calculated proportions.

Folder Information:

  • /preprocessing/: notebook to process data and create pseudobulks for experiments.
  • /exploratory_experiments/: notebooks containing the deconvolution and factorization of different types of pseudobulks.
  • /functions/: scripts with functions used throughout other folders.
  • /data/: contains data information. See Data Information below.

Data information:

Note:

  • CIBERSORTx was run through Docker with user-specific code using the exploratory_experiments/Run_cibersortx.py script.
  • Codes can be given through the CIBERSORTx team by contacting them through the CIBERSORTx website.

The methodology can be roughly divided into experiment 1, 2, 3 and 4. The following outline describes each file and the rough outline in which the scripts can be run.

EXP1: NNLS deconvolution of pseudobulks with distinct immune cell-types.

  • /preprocessing/EXP1_pseudos_snadp.ipynb
    • Preprocess and creates pseudobulks from SN adipose tissue data.
  • /exploratory_experiments/EXP1.py
    • Runs EXP1_eval.ipynb through Papermill. Parameters must be specific in EXP1.py file.
  • /exploratory_experiments/EXP1_eval.ipynb
    • NNLS deconvolution and residual analysis.

EXP2: NNLS, BayesPrism, and CIBERSORTx deconvolution of pseudobulks from PBMC3k.

  • /preprocessing/EXP2_pseudos_pbmc.ipynb
    • Preprocess and creates pseudobulks from PBMC3k 10x Genomics dataset.
  • /preprocessing/EXP2_ciber_bayes_prep.ipynb
    • Creates files needed to run BayesPrism and CIBERSORTx.
  • /preprocessing/EXP2_pbmc_BayesPrism.R
    • R script for BayesPrism deconvolution of files created in EXP2_ciber_bayes_prep.ipynb.
  • /exploratory_experiments/EXP2.py
    • Runs EXP2_eval.ipynb through Papermill. Parameters must be specific in EXP2.py file.
  • /exploratory_experiments/EXP2_cibersort_eval.ipynb
    • Analyzes the results of CIBERSORTx deconvolution.
  • /exploratory_experiments/EXP2_bayesprism_eval.ipynb
    • Analyzes the results of BayesPrism deconvolution.
  • /exploratory_experiments/EXP2.py
    • Runs EXP2_eval.ipynb through Papermill. Parameters must be specific in EXP2.py file.
    • This should be run after both BayesPrism and CIBERSORTx are done.
  • /exploratory_experiments/EXP2_eval.ipynb
    • Final notebook containing the NNLS deconvolution and analysis, and the comparison between deconvolution methods.

EXP3: NNLS, BayesPrism, and CIBERSORTx deconvolution of pseudobulks from single-nucleus adipose tissue, with single-cell RNA-seq missing real cell-types. Incorporated noise and realistic proportions.

  • /preprocessing/EXP3_pseudos_snadp.ipynb
    • Preprocess and creates pseudobulks from white adipose tissue of single-nucleus and single-cell data.
  • /preprocessing/EXP3_ciber_bayes_prep.ipynb
    • Creates files needed to run BayesPrism and CIBERSORTx.
  • /preprocessing/EXP3_adp_BayesPrism.R
    • R script for BayesPrism deconvolution of files created in EXP3_ciber_bayes_prep.ipynb.
  • /exploratory_experiments/EXP3_cibersort_eval.ipynb
    • Analyzes the results of CIBERSORTx deconvolution.
  • /exploratory_experiments/EXP2_bayesprism_eval.ipynb
    • Analyzes the results of BayesPrism deconvolution.
  • /exploratory_experiments/EXP3.py
    • Runs EXP3_eval.ipynb through Papermill. Parameters must be specific in EXP3.py file.
    • This should be run after both BayesPrism and CIBERSORTx are done.
  • /exploratory_experiments/EXP3_eval.ipynb
    • Final notebook containing the NNLS deconvolution and analysis, and the comparison between deconvolution methods.

EXP4: Real bulks of HGSOC deconvolved with NNLS (See Discussion)

  • /preprocessing/EXP4_preprocessing.ipynb
    • Preprocess and QCs single-cell and bulk data from HGSOC.
  • /exploratory_experiments/EXP4_eval.ipynb
    • NNLS deconvolution and analysis of dissociated and classic bulks.

Setting up the Conda Environment

To run the experiments in this repository, you'll need to set up a Conda environment named env_ml. The required configuration file, env_ml.yml, is located in the "environment" folder of this repository.

Prerequisites

Make sure you have Conda installed. If not, you can download and install it from Conda's official website.

Creating the Conda Environment

Navigate to the root of the repository and run the following command to create the env_ml environment:

conda env create -f environment/env_ml.yml

About

Predicting missing cell-type proportions from deconvolution residual using NNLS.

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages