-
Notifications
You must be signed in to change notification settings - Fork 0
Commit
This commit does not belong to any branch on this repository, and may belong to a fork outside of the repository.
- Loading branch information
1 parent
3560809
commit 4e9c315
Showing
19 changed files
with
579 additions
and
74 deletions.
There are no files selected for viewing
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
| Original file line number | Diff line number | Diff line change |
|---|---|---|
| @@ -0,0 +1,236 @@ | ||
| # Define global variables | ||
| utils::globalVariables(c("alpha")) | ||
|
|
||
| #' Obtain box plots for the correlations of all celltypes, and the mean correlations at a specified cutoff p-value | ||
|
|
||
| #' @importFrom ggplot2 geom_boxplot ggplot labs facet_wrap theme aes element_text element_blank scale_colour_brewer scale_alpha guides guide_legend | ||
| #' @importFrom cowplot theme_cowplot | ||
|
|
||
| #' @param corrMats (named) list of correlation matrices for each celltype with the final element being the mean correlation matrix, all at specified p-value | ||
| #' @param numRealDatasets total number of *real* datasets (most likely the number of studies, but sometimes a study may be split e.g. into 2 brain regions, so in this case it would be the number of studies plus 1) | ||
| #' @param pval the cut-off p-value which was used to select DEGs | ||
| #' @param alphaval (alpha) transparency of the non-mean boxplots | ||
| #' @param numPerms number of random permutations of the dataset used to select significant DEGs from | ||
| #' @param numSubsets number of pairs of random subsets of the dataset used to select significant DEGs from | ||
| #' @param sexDEGs true if DEGs come from sex chromosomes, else false | ||
| #' @param fontsize_yaxislabels font size for axis labels in plot | ||
| #' @param fontsize_yaxisticks font size for axis tick labels in plot | ||
| #' @param fontsize_title font size for plot title | ||
| #' @param fontsize_legendlabels font size for legend labels in plot | ||
| #' @param fontsize_legendtitle font size for legend title in plot | ||
| #' @param fontsize_facet_labels font size for facet labels | ||
|
|
||
| #' @return box plots for correlation matrices at a certain p-value cut-off, sorted by celltype and then type of correlation | ||
|
|
||
| correlation_boxplots <- function(corrMats, | ||
| numRealDatasets, | ||
| pval, | ||
| alphaval=0.25, | ||
| numPerms=5, | ||
| numSubsets=5, | ||
| sexDEGs=FALSE, | ||
| fontsize_yaxislabels=12, | ||
| fontsize_yaxisticks=9, | ||
| fontsize_title=14, | ||
| fontsize_legendlabels=9, | ||
| fontsize_legendtitle=9, | ||
| fontsize_facet_labels=9){ | ||
|
|
||
| # check input parameters are fine | ||
| if(class(corrMats)!="list"){ | ||
| stop("Error: corrMats should be a list of matrices") | ||
| } | ||
| if(floor(numRealDatasets)!=numRealDatasets){ | ||
| stop("Error: numRealDatasets should be an integer specifying the number of real datasets (see description)") | ||
| } | ||
| if(!is.numeric(pval) | pval <= 0 | pval > 1){ | ||
| stop("Error: pval should be a (positive) number between 0 and 1") | ||
| } | ||
| if(floor(numPerms)!=numPerms){ | ||
| stop("Error: numPerms should be an integer specifying the number of random permutations of Tsai (see description)") | ||
| } | ||
| if(floor(numSubsets)!=numSubsets){ | ||
| stop("Error: numSubsets should be an integer specifying the number of subsets of Tsai (see description)") | ||
| } | ||
| if(class(sexDEGs)!="logical"){ | ||
| stop("Error: sexDEGs should be TRUE (if DEGs are chosen only from sex chromosomes) or FALSE") | ||
| } | ||
| if(class(fontsize_yaxislabels)!="numeric"){ | ||
| stop("Error: fontsize_yaxislabels should be numerical.") | ||
| }else{ | ||
| if(fontsize_yaxislabels-floor(fontsize_yaxislabels)!=0|fontsize_yaxislabels<0){ | ||
| stop("Error: fontsize_yaxislabels should be a positive integer.") | ||
| } | ||
| } | ||
| if(class(fontsize_yaxisticks)!="numeric"){ | ||
| stop("Error: fontsize_yaxisticks should be numerical.") | ||
| }else{ | ||
| if(fontsize_yaxisticks-floor(fontsize_yaxisticks)!=0|fontsize_yaxisticks<0){ | ||
| stop("Error: fontsize_yaxisticks should be a positive integer.") | ||
| } | ||
| } | ||
| if(class(fontsize_title)!="numeric"){ | ||
| stop("Error: fontsize_title should be numerical.") | ||
| }else{ | ||
| if(fontsize_title-floor(fontsize_title)!=0|fontsize_title<0){ | ||
| stop("Error: fontsize_title should be a positive integer.") | ||
| } | ||
| } | ||
| if(class(fontsize_legendlabels)!="numeric"){ | ||
| stop("Error: fontsize_legendlabels should be numerical.") | ||
| }else{ | ||
| if(fontsize_legendlabels-floor(fontsize_legendlabels)!=0|fontsize_legendlabels<0){ | ||
| stop("Error: fontsize_legendlabels should be a positive integer.") | ||
| } | ||
| } | ||
| if(class(fontsize_legendtitle)!="numeric"){ | ||
| stop("Error: fontsize_legendtitle should be numerical.") | ||
| }else{ | ||
| if(fontsize_legendtitle-floor(fontsize_legendtitle)!=0|fontsize_legendtitle<0){ | ||
| stop("Error: fontsize_legendtitle should be a positive integer.") | ||
| } | ||
| } | ||
| if(class(fontsize_facet_labels)!="numeric"){ | ||
| stop("Error: fontsize_facet_labels should be numerical.") | ||
| }else{ | ||
| if(fontsize_facet_labels-floor(fontsize_facet_labels)!=0|fontsize_facet_labels<0){ | ||
| stop("Error: fontsize_facet_labels should be a positive integer.") | ||
| } | ||
| } | ||
|
|
||
| # midCor submatrix limits | ||
| midCorLim <- numPerms + numRealDatasets | ||
| # list to hold results | ||
| corrOuts <- c() | ||
| # index | ||
| j <- 1 | ||
|
|
||
| # get lists with all correlations | ||
| for(corrMat in corrMats){ | ||
| # specify submatrices with upper/middle/lower bounds | ||
| lower <- corrMat[1:numPerms,1:numPerms] | ||
| middle <- corrMat[(numPerms+1):midCorLim,(numPerms+1):midCorLim] | ||
| upper <- corrMat[(midCorLim+1):(midCorLim+numSubsets),(midCorLim+1):(midCorLim+numSubsets)] | ||
| # convert each one to a list and remove "1" (selfcorrelation) | ||
| lower <- unique(unlist(as.list(lower))) | ||
| lower <- lower[-c(1)] | ||
| middle <- unique(unlist(as.list(middle))) | ||
| middle <- middle[-c(1)] | ||
| upper <- unique(unlist(as.list(upper))) | ||
| upper <- upper[-c(1)] | ||
| # store in list | ||
| corrOuts[[j]] <- list(lower,middle,upper) | ||
| names(corrOuts)[[j]] <- names(corrMats)[[j]] | ||
| # increment | ||
| j <- j+1 | ||
| } | ||
|
|
||
| # store in dataframe | ||
| i <- 1 | ||
| # empty dataframe | ||
| df <- data.frame() | ||
| # fill dataframe | ||
| for(out in corrOuts){ | ||
| # define variables | ||
| var1 <- replicate(length(out[[1]])+length(out[[2]])+length(out[[3]]), names(corrOuts)[[i]]) #celltype | ||
| var2 <- c(replicate(length(out[[1]]),"Random Permutations"),replicate(length(out[[2]]),"Between Study"),replicate(length(out[[3]]),"Within-study subsamples")) | ||
| val <- unlist(out) | ||
| # put in dataframe | ||
| df_new <- data.frame(var1) | ||
| df_new$var2 <- var2 | ||
| df_new$val <- val | ||
| # join | ||
| df <- rbind(df,df_new) | ||
| i <- i+1 | ||
| } | ||
|
|
||
| df$alpha <- replicate(dim(df)[[1]],alphaval) | ||
| df$alpha <- ifelse(df$var1=="Mean", 1, ifelse(df$var1!="Mean",alphaval,alphaval)) | ||
| unique_alphas <- df[!duplicated(df[,c("var1")]),]$alpha | ||
|
|
||
| # box plot | ||
| if(pval == 1){ | ||
| if(sexDEGs == FALSE){ | ||
| fig.plot <- ggplot(df, | ||
| aes(x=factor(var1,levels=c("Mean","Astro","Endo","Micro","Oligo")),y=val))+ | ||
| geom_boxplot(outlier.shape=NA,aes(fill=factor(var1,levels=c("Mean","Astro","Endo","Micro","Oligo")),alpha=alpha),width=5)+ | ||
| theme_cowplot()+ | ||
| scale_colour_brewer(palette = "Set1")+ | ||
| labs(y="Correlation", x="Type", fill="Cell Type",title=paste0("Using all DEGs (from all chromosomes)"))+ | ||
| facet_wrap(factor(df$var2,levels=c("Random Permutations","Between Study","Within-study subsamples")), scales="fixed")+ | ||
| theme(axis.title.x=element_blank(), | ||
| axis.text.x=element_blank(), | ||
| axis.ticks.x=element_blank(), | ||
| axis.title.y = element_text(size = fontsize_yaxislabels), | ||
| axis.text.y = element_text(size = fontsize_yaxisticks), | ||
| plot.title = element_text(size = fontsize_title), | ||
| legend.text = element_text(size = fontsize_legendlabels), | ||
| legend.title = element_text(size = fontsize_legendtitle), | ||
| strip.text = element_text(size = fontsize_facet_labels))+ | ||
| scale_alpha(guide = 'none')+ | ||
| guides(fill=guide_legend(override.aes = list(alpha=unique_alphas))) | ||
| }else{ | ||
| fig.plot <- ggplot(df, | ||
| aes(x=factor(var1,levels=c("Mean","Astro","Endo","Micro","Oligo")),y=val))+ | ||
| geom_boxplot(outlier.shape=NA,aes(fill=factor(var1,levels=c("Mean","Astro","Endo","Micro","Oligo")),alpha=alpha),width=5)+ | ||
| theme_cowplot()+ | ||
| scale_colour_brewer(palette = "Set1")+ | ||
| labs(y="Correlation", x="Type", fill="Cell Type",title=paste0("Using all DEGs (from sex chromosomes)"))+ | ||
| facet_wrap(factor(df$var2,levels=c("Random Permutations","Between Study","Within-study subsamples")), scales="fixed")+ | ||
| theme(axis.title.x=element_blank(), | ||
| axis.text.x=element_blank(), | ||
| axis.ticks.x=element_blank(), | ||
| axis.title.y = element_text(size = fontsize_yaxislabels), | ||
| axis.text.y = element_text(size = fontsize_yaxisticks), | ||
| plot.title = element_text(size = fontsize_title), | ||
| legend.text = element_text(size = fontsize_legendlabels), | ||
| legend.title = element_text(size = fontsize_legendtitle), | ||
| strip.text = element_text(size = fontsize_facet_labels))+ | ||
| scale_alpha(guide = 'none')+ | ||
| guides(fill=guide_legend(override.aes = list(alpha=unique_alphas))) | ||
| } | ||
| }else{ | ||
| if(sexDEGs == FALSE){ | ||
| fig.plot <- ggplot(df, | ||
| aes(x=factor(var1,levels=c("Mean","Astro","Endo","Micro","Oligo")),y=val))+ | ||
| geom_boxplot(outlier.shape=NA,aes(fill=factor(var1,levels=c("Mean","Astro","Endo","Micro","Oligo")),alpha=alpha),width=5)+ | ||
| theme_cowplot()+ | ||
| scale_colour_brewer(palette = "Set1")+ | ||
| labs(y="Correlation", x="Type", fill="Cell Type",title=paste0("DEGs selected at a ",pval*100,"% cut-off (from all chromosomes)"))+ | ||
| facet_wrap(factor(df$var2,levels=c("Random Permutations","Between Study","Within-study subsamples")), scales="fixed")+ | ||
| theme(axis.title.x=element_blank(), | ||
| axis.text.x=element_blank(), | ||
| axis.ticks.x=element_blank(), | ||
| axis.title.y = element_text(size = fontsize_yaxislabels), | ||
| axis.text.y = element_text(size = fontsize_yaxisticks), | ||
| plot.title = element_text(size = fontsize_title), | ||
| legend.text = element_text(size = fontsize_legendlabels), | ||
| legend.title = element_text(size = fontsize_legendtitle), | ||
| strip.text = element_text(size = fontsize_facet_labels))+ | ||
| scale_alpha(guide = 'none')+ | ||
| guides(fill=guide_legend(override.aes = list(alpha=unique_alphas))) | ||
| }else{ | ||
| fig.plot <- ggplot(df, | ||
| aes(x=factor(var1,levels=c("Mean","Astro","Endo","Micro","Oligo")),y=val))+ | ||
| geom_boxplot(outlier.shape=NA,aes(fill=factor(var1,levels=c("Mean","Astro","Endo","Micro","Oligo")),alpha=alpha),width=5)+ | ||
| theme_cowplot()+ | ||
| scale_colour_brewer(palette = "Set1")+ | ||
| labs(y="Correlation", x="Type", fill="Cell Type",title=paste0("DEGs selected at a ",pval*100,"% cut-off (from sex chromosomes)"))+ | ||
| facet_wrap(factor(df$var2,levels=c("Random Permutations","Between Study","Within-study subsamples")), scales="fixed")+ | ||
| theme(axis.title.x=element_blank(), | ||
| axis.text.x=element_blank(), | ||
| axis.ticks.x=element_blank(), | ||
| axis.title.y = element_text(size = fontsize_yaxislabels), | ||
| axis.text.y = element_text(size = fontsize_yaxisticks), | ||
| plot.title = element_text(size = fontsize_title), | ||
| legend.text = element_text(size = fontsize_legendlabels), | ||
| legend.title = element_text(size = fontsize_legendtitle), | ||
| strip.text = element_text(size = fontsize_facet_labels))+ | ||
| scale_alpha(guide = 'none')+ | ||
| guides(fill=guide_legend(override.aes = list(alpha=unique_alphas))) | ||
| } | ||
| } | ||
|
|
||
| return(fig.plot) | ||
|
|
||
| } |
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Oops, something went wrong.