BronchiCellPred

To predict bronchial cell proportions in bulk RNA-seq datasets

https://github.com/CancanQi/BronchiCellPred/

Introduction

BronchiCellPred is an r package containing models for predicting the proportions of bronchial epithelial cells using bulk gene expression data from bronchial biopsies.

Gene signature matrix used for cell type deconvolution were generated using a single-cell RNAseq dataset of bronchial biopsies.

Three cell type deconvolution methods were available in this package, including nnls, MuSiC (https://github.com/xuranw/MuSiC) and BSeq-sc(https://github.com/shenorrLab/bseqsc). Users can choose the methods according to their own preference.

For additional details on methods and results please go our [manuscript](link to be updated).

Installation

library(devtools)
install_github("CancanQi/BronchiCellPred")
## or 
install_github("CancanQi/BronchiCellPred",ref="HEAD")

Quick start

Load the library and dependencies.

library(MuSiC)
library(bseqsc)
library(xbioc)
library(reshape2)
library(tidyverse)
library(Biobase)

Prepare input data

## read bulk gene expression count table, with genes in row and samples in column
count.table<-read.table("your file name")
## convert count table to expression set
bulk.eset <- Biobase::ExpressionSet(assayData = data.matrix(count.table))

Predict bronchial cell proportions

## use the bulk expression set as the input, chose the methods from nnls, MuSiC and bseq
est_prop<-BronCell.prop(bulk.eset,method="MuSiC")
## the output is a dataframe of cell proporion for each cell type (column) and each sample (row)

## plot the results
BronCell.plot(est_prop)