Growth curves in R with standard deviation - ggplot2

Question

I am trying to plot my data (replicate results for each strain) and i want only one line graph for each strain, this means averaged results of replicates for each strain with points along the line with error bars (error between replicate data).
If you click on the image above, it shows the plot i have so far, which displays WT and WT.1 as seperate lines and all other replicates. However, they are replicates of each strain (WT,DrsbR,DsigB) and i want them to appear as one line of mean results for each strain instead. I am using ggplot package- and melting data with reshape package, but cannot figure out how to make my replicates appear as one line together with error bars (standard deviation of mean results between replicates).
The image in black and white is something i am looking for in my graph- seperate line with points of replicate data plotted as a mean value.
library(reshape2)
melted<-melt(abs2)
print(abs2)
melted<-melt(abs2,id=1,measured=c("WT","WT.1","DsigB","DsigB.1","DrsbR","DrsbR.1"))
View(melted)
colnames(melted)<-c("Time","Strain","Values")
##line graph for melted data
melted$Time<-as.factor(melted$Time)
abs2line=ggplot(melted,aes(Time,Values))+geom_line(aes(colour=Strain,group=Strain))
abs2line+
stat_summary(fun=mean,
geom="point",
aes(group=Time))+
stat_summary(fun.data=mean_cl_boot,
geom="errorbar",
width=.2)+
xlab("Time")+
ylab("OD600")+
theme_classic()+
labs(title="Growth Curve of Mutant Strains")
summary(melted)
print(melted)

Answer 1

One approach is to take your melted data frame and separate out the "variable" column into "species" and "strain" using the separate() function from tidyr. I don't have your dataset -- it is appreciated if you are able to share your dataset via dput(your.data.frame) for future questions -- so I made a dummy dataset that's similar to yours. Here we have two "species" (red and blue) and two "strains" for each species.
df <- data.frame(
time = seq(0, 40, by=10),
blue = c(0:4),
blue.1 = c(0, 1.1, 1.9, 3.1, 4.1),
red = seq(0, 8, by=2),
red.1 = c(0, 2.1, 4.2, 5.5, 8.2)
)
df.melt <- melt(df,
id.vars = 'time',
measure.vars = c('blue', 'blue.1', 'red', 'red.1'))
We can then use tidyr::separate() to separate the resulting "variable" column into a "species" column and a "strain" column. Luckily, your data contains a "." which can be a handy character to use for the separation:
df.melt.mod <- df.melt %>%
separate(col=variable, into=c('species', 'strain'), sep='\\.')
Note: The above code will give you a warning related to the point that "blue" and "red" do not have the "." character, thereby giving you NA for the "strain" column. We don't care here, because we're not using that column for anything here. In your own dataset, you can similarly not care too much.
Then, you can actually just use stat_summary() for all geoms... modify as you see fit for your own visual and thematic preference. Note that order matters for layering, so I plot geom_line first, then geom_point, then geom_errorbar. Also note that you can assign the group=species aesthetic in the base ggplot() call and that mapping applies to all geoms unless overwritten.
ggplot(df.melt.mod, aes(x=time, y=value, group=species)) +
stat_summary(
fun = mean,
geom='line',
aes(color=species)) +
stat_summary(
fun=mean,
geom='point') +
stat_summary(
fun.data=mean_cl_boot,
geom='errorbar',
width=0.5) +
theme_bw()