Scatter Plot for two sample condition

I have data set between two comparison or two condition such as HSC and CMP , so i want to do a scatter plot with along with the regression value. Meanwhile when i m trying to label the sample they are all labelled into same color , i want different color for the labelling so that it can be distinguished

Here is my sample data set

               gene        HSC       CMP
     ENSG00000158292.6  1.8102636  2.456869
     ENSG00000162496.6  2.6796705  6.203838
    ENSG00000117115.10  3.4509115  5.555739
    ENSG00000159423.14  3.6809277  5.063446
     ENSG00000053372.4  5.7089974  6.851090
     ENSG00000127423.8  4.4894292  5.996304
     ENSG00000242125.3 10.6258802 11.715932

Here is my code

library("ggpubr")
ggscatter(data1, x = "HSC", y = "CMP", 
          add = "reg.line", conf.int = FALSE, 
          cor.coef = TRUE, cor.method = "pearson",
          color = "black", size=1)

Any suggestion or help would be highly appreciated

• 4,697 views • link

updated 8.7 years ago by Nicolas Rosewick

• written 8.7 years ago by 1769mkc

1 answer

If I understand well you want a scatterplot with a linear regression ?

Edit to take your comment into account i.e. the color variable

using ggplot if data1 is your dataframe :

    ggplot(data1,aes(x=HSC,y=CMP,col=variable)) +
        geom_point() +
        geom_smooth(method = "lm", se = TRUE)

• 0 views • link

written 8.7 years ago by Nicolas Rosewick

library(reshape2)

ex <- melt(gendat, id.vars=c("HSC", "CMP"))
ggplot(ex, aes(x=HSC, y=CMP, color=variable)) + 
  geom_point(shape = 20) +
  geom_smooth(method="lm", col = "black",se = TRUE)

I tried that too...

• 0 views • link

written 8.7 years ago by 1769mkc

what is the variabl "variable" ? I do not see it in your original question.

• 0 views • link

written 8.7 years ago by Nicolas Rosewick

The variable is after i melt the data

       HSC       CMP      variable              value
    1.8102636  2.456869  gene               ENSG00000158292.6
     2.6796705  6.203838  gene              ENSG00000162496.6
     3.4509115  5.555739  gene              ENSG00000117115.10

• 0 views • link

written 8.7 years ago by 1769mkc

I edited my answer. So each point will be colored relative to the "variable" variable.

• 0 views • link

written 8.7 years ago by Nicolas Rosewick

Or am I melting the data in a wrong way perhaps ?

• 0 views • link

written 8.7 years ago by 1769mkc

Maybe you question is not clear. You want a scatterplot where each point is colored based on a variable (in this case defined in the "variable" column). If yes then my answer should work.

• 0 views • link

written 8.7 years ago by Nicolas Rosewick

so i wanted that if lets say i have a sample called HSC so all the genes which are from HSC should have different color and same with the other sample .I second code i posted is different from the first one i get the variable column after I melt the dataframe .

• 0 views • link

written 8.7 years ago by 1769mkc

I thought all genes were found in the two samples ? You could add an additional column to check if the gene is found in the sample of interest (so > 0).

data1$inHSC <- ifelse(data1$HSC>0,TRUE,FALSE)

and then

ggplot(data1,aes(x=HSC,y=CMP,col=inHSC)) + geom_point() + geom_smooth(method = "lm", se = TRUE)

• 0 views • link

written 8.7 years ago by Nicolas Rosewick

yes the genes are common to both the samples .

• 0 views • link

written 8.7 years ago by 1769mkc

ok so the answer in my comment just above is ok for you ?

• 0 views • link

written 8.7 years ago by Nicolas Rosewick

yes it works ,is there a way to display R^2 and to add to it lets say i want to label my data point based on the gene names after certain threshold ? For example if the difference between two point is > 1 or 2 I want highlight or label the point

• 0 views • link

written 8.7 years ago by 1769mkc