GCAncestors contains the Matlab code for uncovering large networks of dependency between time series. It is an iterative procedure based on Granger Causality that can be very easily parallelized. It is currently being ported to R.
The method is described and used in our BMC article Identifying interactions in the time and frequency domains in local and global networks – A Granger Causality Approach. Our first application on real data was on a proteomic dataset of 800 time series. The final network looks as follows:
Download the PDF to zoom in and out of this image.