Perform computational analysis from Ganio et al. looking at the effects of steroids on the immune system after surgery. We used the VoPo clustering algorithm (https://github.com/stanleyn/VoPo) to generate immune features from the 331 samples available on flow repository (https://flowrepository.org/id/FR-FCM-Z2AT).
May 11, 2020
Please contact Natalie Stanley (NatalieStanley1318@gmail.com) with any questions
-
This code was implemented and tested in R version 3.4.4 and uses randomForest version 4.6-14.
-
Before getting started, please make sure you have the following dependencies installed in R (
Biobase,randomForest,pROC,plyr,foreach,doParallel,ggplot2,reshape2,plyr) -
Clone this repository
git clone https://github.com/stanleyn/steroid_immune_data- You have cloned this directory and into some location YourPath. Start R, and change into the YourPath directory
> setwd('YourPath')-You can use the script RF_PerTP.R to create a random forest model for each individual timepoint
> source('RF_PerTP.R')-
When this is finished running, you can see the vector of AUCs and Wilcoxon p-values corresponding to the predictive power of the model learned for each timepoint (Pre, 1hr, 6hr, 24r, 48hr, and 2 wks after surgery)
-
To see AUCs,
> AUC- To see Wilcoxon p-values,
> Wilcox- A plot of boxplots per timepoint will be generated to the main directory called
steroid_classif.pdf