Problems with relative abundance graphs

[md.genmicro]

Hello everybody,

I´m struggling to fix the scale of my relative abundance graph with ggplot2 for a phyloseq class object. 

The script I´m using is:

#Calculating relative abundance for the whole set and selecting top5 classes:

dataf

datag = tax_glom(dataf, "Rank4") 

datarg = transform_sample_counts(datag, function(x) x/sum(x)) 

class.sumf = tapply(taxa_sums(datarg), tax_table(datarg)[, "Rank3"], sum, na.rm = TRUE)

top5fclass = names(sort(class.sumf, TRUE))[1:5]

top5f = prune_taxa((tax_table(datarg)[, "Rank3"] %in% top5fclass), datarg)

#Subsetting samples (I have 4 groups)

groupf1 = subset_samples(top5f, Group == "1" )

groupf1m = merge_samples(groupf1, group = "Date", fun=mean)

pf1 = plot_bar(groupf1m, "Rank3", fill = "Rank3", facet_grid =~~ sample_names(groupf1m), title = "X1" ) + scale_fill_brewer(palette = "Set2") + guides(fill=FALSE) + xlab("Class") + ylab("Relative Abundance") + guides(fill=FALSE)

pf1 = pf1 + geom_bar(stat="identity") 

pf1  

#Multiple plot

require(cowplot)

theme_set(theme_cowplot(font_size=8))

pf = plot_grid(pf1, pf2, pf3, pf4, ncol = 2, align = 'h')

pf 

#Results:

 My result is a multiple grid, with different scales, with numbers that doesn´t make much sense for me (like from 0 to 4 in one grid, the other from 0 to 2)... when I try to put everything in the same scale with ylim, some bars disappear... besides that, the sum of all bars (for each group) is larger than 1. 

If anyone have an idea of what I can do to fix, I ´ll appreciate very much! If you have any comment to improve this script (I´m a beginner), it´s also great!

Thanks in advance!