PlotGroups - heatmap of p-values

Hi,
Here is my proposal for code

Lines 1606 to 1616 in af08bb6

    
           score <- rep(NA,ncol(values)) 
        
           for(i in seq_len(ncol(values))){ 
        
               test <- stats::wilcox.test(values[groups$groups %in%  
        
                                       groupnames[1],i], 
        
                                   values[groups$groups %in% group,i]) 
        
               #boxplot(counts[1:5,i],counts[11:15,i],main=test$p.value) 
        
               adj_p <- stats::p.adjust(test$p.value,"BH") 
        
               diff <- groups$means[group,i] -  
        
                   groups$means[groupnames[1],i] 
        
               score[i] <- adj_p * (-1)^(diff < 0) 
        
           }

I applied same correction as before (although not rigorous). I turned the score in the minus the log10 of the p-value (which is standard in volcano plot). I adapted the colour scale. I set the test as not exact in order to avoid warnings with tied values. Once again, this is just a proposal. I would prefer to get a symmetrical color scale, but this requires more job.
Cheers.

        p_v <- rep(NA,ncol(values))
        for(i in seq_len(ncol(values))){
          test <- stats::wilcox.test(
            values[groups$groups %in% groupnames[1],i],
            values[groups$groups %in% group,i],
            exact = FALSE)
          p_v[i] <- test$p.value
        }
        diff <- groups$means[group,] - groups$means[groupnames[1],]
        adj_p <- stats::p.adjust(p_v,"BH")
        score <- -log10(adj_p) * (-1)^(diff < 0)

	score <- rep(NA,ncol(values))
	for(i in seq_len(ncol(values))){
	test <- stats::wilcox.test(values[groups$groups %in%
	groupnames[1],i],
	values[groups$groups %in% group,i])
	#boxplot(counts[1:5,i],counts[11:15,i],main=test$p.value)
	adj_p <- stats::p.adjust(test$p.value,"BH")
	diff <- groups$means[group,i] -
	groups$means[groupnames[1],i]
	score[i] <- adj_p * (-1)^(diff < 0)
	}