In genomic papers the most interesting stuff is usually hiding in supplements.
Lazaridis discuss the origins of Armenian language.
He mention the remarkable homogeneity of Lchashen Metsamor culture he propose that they could be the Armenian speakers.
Then later in page 280 he discuss the dilution of Steppe ancestry ( which by the way is related but NOT equal to EHG ) in post Urartian period he speculates that this data opens the possibility of migration from Balkans but few sentences later acknowledge that there is virtually no evidence of such a migration. And says that the direct migration from North is the most realistic.
In reality this dilution is not a big problem for a theory from North. Myceneans have the same dilution. Even more in some cases.
Nothing forbids us to imagine that some groups moved further south from modern RoA and settled in Van later returned back with Urartu conquest. Also there is no such a thing as post-Urartian large migrations. The true Urartian sites in RoA are not sampled in this paper. And they obviously had a large population from South. So those low Steppe people came in Urartian period. Maybe some came later. But the bulk is from Urartian period. So if this dilution has something to do with the introduction of Armenian language then one must agree that Urartu was in some part Armenian. But as You understand this theory has some problems of linguistic nature. That's probably is the reason why they do not discuss such a possibility.
In sum for better understanding this dilution of Steppe ancestry it is very important to have dense sampling from Urartian sites in RoA. Numerous bones exist from those sites so there is no technical obstacles.