Question: Trendline Of Chipseq Read Values In Intron Regions
1
gravatar for daniel.soronellas
8.1 years ago by
Spain
daniel.soronellas330 wrote:

Hi everyone,

I post today to ask for help in order to use ggplot2 package, but first the fundamental question:

Is possible to join in one plot multiple trendlines of different datasets?

The reason of this question is because I need to visualize in the same plot, the ratio of reads between 3 chipseq samples and 1 input using a trendline for each sample in a big intronic region. So far I've tried to do this with GenomeGraph R package but it was confusing for me to get this plot.

So I decided to do it with ggplot2 package. The input is a file that have almost 300 bins of this big intron and for each bin the number of reads of each sample:

iD          pol.t0   pol.t5  pol.t60  input
Intron279    11    10    5    17
Intron278    10    24    20    30
Intron277    17    58    20    41
Intron276    5    29    9    38
Intron275    5    16    6    29
Intron274    4    14    10    28
Intron273    11    33    17    28
Intron272    11    40    15    36
Intron271    8    18    4    36
Intron270    7    24    5    32

And this is my Rscript to plot the final trends in a separate plot each:

library(ggplot2)

t<-read.table("BigIntron_reordered.bed")

## Ratio Calculation
t2<-data.frame(t[,1],  (t[,2]/t[,5]), (t[,3]/t[,5]), (t[,4]/t[,5]))
colnames(t2)<-c("iD","pol.t0","pol.t5","pol.t60")

## No lose initial order of iDs
t2$iD<-factor(t2$iD, levels=unique(as.character(t2$iD)) )   

## Plot trendlines for each sample
p1<-ggplot(t2[t2$pol.t0 != "Inf" & t2$pol.t0 != "NaN",], aes(x=iD, y=pol.t0)) + stat_smooth(method="loess",aes(group=1),size=2)
p2<-ggplot(t2[t2$pol.t0 != "Inf" & t2$pol.t0 != "NaN",], aes(x=iD, y=pol.t5)) + stat_smooth(method="loess",aes(group=1),size=2) 
p3<-ggplot(t2[t2$pol.t0 != "Inf" & t2$pol.t0 != "NaN",], aes(x=iD, y=pol.t60)) + stat_smooth(method="loess",aes(group=1),size=2)

## Grid plot of everything together
grid.arrange(p1, p2, p3, nrow=3)

The final result is:

So is possible to join this trend lines all together in one plot?

Thanks for your help!!

R intron chipseq • 2.5k views
ADD COMMENTlink modified 3.8 years ago by Biostar ♦♦ 20 • written 8.1 years ago by daniel.soronellas330
3
gravatar for Sebastian Kurscheid
8.1 years ago by
Australia, ACT, Canberra, ANU
Sebastian Kurscheid300 wrote:

Try the following - it's not quite what you wanted (intron label on the x-axis is lost), but it's all three curves in one plot)

First transform the dataframe:

require(reshape)
t2.1 <- melt(t2, id = "iD", variable_name = "series")
t2.1$iD <- rep(1:10,3) #Apparently the subsequent qplot() function does not like the characters in the iD column

Then plot it:

ggplot(t2.1, aes(iD, value)) + geom_smooth(method="loess", aes(colour = series))

enter image description here

Self-Edit: Comment: I just used the lines of data provided in your question, I suppose that explains the different shapes of the curves?

ADD COMMENTlink modified 8.1 years ago • written 8.1 years ago by Sebastian Kurscheid300

Yes, sure it is for that! Thanks for your answer was perfectly clear!

ADD REPLYlink written 8.1 years ago by daniel.soronellas330
Please log in to add an answer.

Help
Access

Use of this site constitutes acceptance of our User Agreement and Privacy Policy.
Powered by Biostar version 2.3.0
Traffic: 2893 users visited in the last hour
_