I realize you've kind of already covered this territory, but I've read through the entire tree of comments here and I feel really uncertain about which files I should include for an "optimal" combination. I have at least some version of all the IWADs in question (there are multiple versions of almost all of them) but I'm not sure exactly which version of each IWAD is most desirable to create the best version of the output.
Would it be possible to create an "ideal IWAD list", based on the checksum information from these sites?
To minimize the work required on your end (which I certainly want to do), it seems like the most logical process would be for you to collect the ideal IWAD collection in a folder on your own system and then just run sha1sum (or md5sum, etc.) against them and post the checksums. Then I can easily figure out exactly which version of each IWAD those checksums correlate to (I've already written a tool to do this, as preparation for writing this comment).