Contrast matrices created by Hypothesis have coefficients as rownames and each column is a contrast. Inside getLogFC, if no contrast 0 is supplied, one is created as a single row matrix with coefficients as columns. Contrast 0 and 1 are then supposed to be rbinded, but the dimensions are not equal. getLogFC fails with the message
Error in rbind(...) : number of columns of matrices must match (see arg 2)
It seems like within the contrast0 check of getLogFC, if contrast0 is supplied as a hypothesis, there should be an additional line to transpose contrast 0 so it's consistent with contrast1 and can be rbound. Something in this area here.
I'm happy to make a pull request if you think this is the proper solution and this is not an intended outcome.
Contrast matrices created by
Hypothesishave coefficients as rownames and each column is a contrast. InsidegetLogFC, if no contrast 0 is supplied, one is created as a single row matrix with coefficients as columns. Contrast 0 and 1 are then supposed to berbinded, but the dimensions are not equal.getLogFCfails with the messageError in rbind(...) : number of columns of matrices must match (see arg 2)It seems like within the contrast0 check of
getLogFC, if contrast0 is supplied as a hypothesis, there should be an additional line to transpose contrast 0 so it's consistent with contrast1 and can be rbound. Something in this area here.I'm happy to make a pull request if you think this is the proper solution and this is not an intended outcome.