linear regression with gene labelling

Question

[1769mkc]

I have a salmonella treated vs KD condition , i want something like this but I not sure how linear regression can be done

Gene    SCRSAL1 KDSAL3
CXCL1   10.8273907686   11.2751328718
CXCL2   8.001064472 8.5753884341
CXCL3   8.1965832765    8.7086794002
CXCL5   8.0973357585    8.6242736829
CXCL6   9.9053940183    10.8940613053
CXCL8   13.1127454083   13.4402538026
CXCL10  7.2590758038    7.8202130661
CCL20   9.0933584064    9.9218889828
CCL22   7.3643976273    7.7911788897
CCL28   8.3723507486    8.6786344048
IL36G   4.3423919715    4.5673548181

This is my dataset the kind of figure i m trying is this Any suggestion or help would be highly appreciated

[Jean-Karim Heriche]

This is not clear at all to me. What is the question ? What are you trying to achieve ? What's the data ?

[russhh]

There only appears to be two columns of data, to talk of multiple regression doesn't make a lot of sense to me

[1769mkc]

oh sorry only one regression it was my typo error...I will fix it

[1769mkc]

I want a scatter plot with regression line and i can label my gene of interest between the comparison, I want to show when there is infection in the scramble as compared to knockdown there is upregualtion of particular set of gene , so can i show that through a regression ?that is my idea..

Answer 1

[Jean-Karim Heriche]

With perfect data, if you plot gene levels for knockdown vs control, any gene above the 45˚ line is upregulated in the knockdown compared to control, the farther away from the line, the stronger the effect. A linear regression will help in dealing with imperfect data. To help decide on which genes are upregulated, you can calculate (and plot) a confidence interval on each side of the regression line. For how to do this in R, see this Rstudio notebook.

[1769mkc]

thank for the resource i will try and let know if im able to get what i want to