This is a test version of Biostars. For the public version, visit https://www.biostars.org.
row t test with biocLite(genefilter)

I'm trying to run a t test on 2 groups of samples for a list of variables. My dataset (=df) looks like this:

GeneName  group1  group1  group1  group2  group2  group2
gene1
gene2
gene3
..

I want to obtain p value for t test comparing for each row (gene) the values of group1 vs group2.

I tried to use the rowttests() function of genefilter package for bioconductor as such:

df$ttest <- rowttests(df, factor(c(group1, group2)))

I'm getting the following error:

Error in (function (classes, fdef, mtable)  :
  unable to find an inherited method for function 'rowttests' for signature '"data.frame", "factor"'

What am I doing wrong?

Thanks

t-test genefilter r
• 6,809 views • link
updated by Ram Ram • written by sebastien200109 • sebastien200109 
df$ttest <- rowttests(as.matrix(df[,-1]), factor(c(0,0,0,1,1,1))

gives the following error:

Error in rowcoltt(x, fac, tstatOnly, 1L) :
  Invalid argument 'x': must be a real matrix.
• 0 views • link
updated by Ram Ram • written by sebastien200109 • sebastien200109

So throw an as.double() in there.

• 0 views • link
written by Devon Ryan Devon Ryan

This is going to get messy, but the general idea is:

df$ttest <- rowttests(matrix(as.double(unlist(df[,-1])), ncol=6), factor(c(0,0,0,1,1,1))

Or something along those lines. BTW, don't worry about the unlist() part, it just converts your data.frame into a vector (data.frames are actually a special kind of list in R).

• 0 views • link
updated by Ram Ram • written by Devon Ryan Devon Ryan

1 answer 

> showMethods(rowttests)
Function: rowttests (package genefilter)
x="ExpressionSet", fac="character"
x="ExpressionSet", fac="factor"
x="ExpressionSet", fac="missing"
x="matrix", fac="factor"
x="matrix", fac="missing"


So, I suspect that something like this will work:

df$ttest <- rowttests(as.matrix(df[,-1]), factor(c(0,0,0,1,1,1))
• 0 views • link
written by Devon Ryan Devon Ryan

Log in to answer this question.

More posts like this