Hi,
I am trying to use the pcalg package to do some causal inference on some
high dimensional omics data (~ 30k variables). It takes forever to run and
my machine get stuck. I just realized that conditional independence test
can be calculated without using the correlation matrix (I think it's just
looping through all the partial correlation calculation), I am wondering
after getting all the edges (maybe in a format of two column matrix with
source and targets in each one), then how to use pc, fci, rfci, etc to do
the causal inference. Right now it seems that we need to supply a
correlation matrix into those algorithms in the sufficient statistics
argument.
Thanks,
-Jack