Is common descent untestable?

[Paying the bill]
Illustration EN&ST uses for "paying the bill", from Wikimedia Commons

The ever-astonishing site Evolution News and Science Today has astonished once again. Finding a paper in the scientific literature which provides evidence for a horizontal gene transfer event, they declare that this shows that common descent is an untestable concept

The paper is in the journal Molecular Biology and Evolution: Citrullination was introduced into animals by horizontal gene transfer from cyanobacteria (Disclosure: I have published in that journal, and served as President of its parent society).

The paper straightforwardly provides evidence from molecular phylogenies and functional studies that the enzyme peptidylarginine deaminase has a distribution discordant with the overall phylogeny of life, in that

The animal and cyanobacterial PADI proteins share functionally relevant primary and tertiary synapomorphic sequences that are distinct from a second PADI type present in fungi and actinobacteria …. The consilience of evidence indicates that PADIs were introduced from cyanobacteria into animals by horizontal gene transfer (HGT).

Aha! exclaims EN&ST. About Universal Common Descent, they note that

UCD is often held to be testable via the counterfactual of “genes and proteins might have been discovered in groups where they should not exist”. In other words, if we observed genes and proteins occurring in the TOL, with unequivocal homology to the same sequences found elsewhere in phylogenetically distant or divergent groups, UCD would be in trouble.

And indeed they think that horizontal gene transfer is an added assumption that needs to be verified, and that if we invoke it in this way, UCD is in trouble:

UCD’s big entourage only becomes a problem when its members always pay the fines for UCD’s failed predictions, pick up UCD’s tab at restaurants and hotels, and generally cover the empirical shortfalls of UCD.

and they end with the declaration that

If ever anyone needed evidence of the untestability of UCD, papers like this provide it.

The even-more-astonishing Denyse O’Leary (“News” at Uncommon Descent) also made the same point, attributing it to “a friend”; one of the two of them also seems to be a friend of EN&ST.

I could argue with these dramatic conclusions, but the commenters here at PT might well do a better job. So let’s hear what you have to say.