gene expression anti-correlation data?

Question

[fuellen]

Hello, I'd like to know, for a set of tissues if possible, whether two genes are anti-correlated (gene A up implying gene B down and vice versa), in NGS datasets and, more generally, in any gene expression datasets. While there are plenty of tools for coexpression (correlation both ways, or just positive correlation), I could not find anything that is anti-correlation only Thanks a lot!

[noorpratap.singh]

Just find the correlation among the genes of interest and if its negative it means they are anti-correlated. Assuming you are using R, you can use the cor function which would return you a g*g matrix where g are the number of genes. Now if there is strong anticorrelation. then you can keep a threshold lets say -0.5 and anything below it gives you the gene pairs. Below is the small code snippet.

cor.df <- cor(df)
apply(cor.df, 1, function(r) which(r < threshold)