I don’t have much criticism. They didn’t have data on direct function, so they employed an algorithm for a very crude initial analysis:
Their goal wasn’t to determine the specific functional roles of the mutations they found. This was an initial sweep of the entire polar bear genome that was meant to find genes that are potentially under selective pressures. Their other goal was to tease out the population dynamics of polar bear speciation. They used PolyPhen as a way of guiding future research, and it was never meant to be the final say on the actual causal role of those specific mutations.
As many here understand, not all conclusions are created equal. There are strong conclusions and weak conclusions. It is up to the reader to determine the strength of those conclusions, and Behe failed to do this, at least in my opinion. It isn’t enough to simply repeat what the authors write and act as if it is gospel. The quality of the conclusion matters.