South Cushites don't descend from East Cushites.
They split very early on from Proto-Cushitic ~300 years +or- and had much higher levels of Eurasian Ancestry then Eastern Cushities with the original population likely being 65-70% Eurasian.
Here for the study.
The reason linguists used to group South Cushitic with Eastern Cushitic is that there was so much influence from the latter into the former which then creates a false connection between the two.
Don't mistake me, I don't believe they were east Cushitic, only that they share a branch with east Cushitic. But they are certianly not like north Cushitic. Iraqw is much much more related to Somali, and Oromo then Beja. They share far too many a series of innovations and developments, not to mention cognates to the exclusion of central Cushitic and north Cushitic. Linguistics is more a science than you take it to be, and does not suffer from the inane arbitrariness of the humanities. Want evidence? look at how the proto-Indo-Europeans were constructed merely from the languages of it's descendants thousands of years later. Their entire society, where they lived, their social relationships, their technologies, subsistence patterns, social contracts, and kinship systems. With the advent of systemic archeology and ancient DNA, and linguistics was proven right over and over again, down to the fact that they were a patriarchal society that revolved around patrilineal kinship. Just take a look at all those Beaker, Afansievo, Yamnaya, and Corded Ware finds, all the product of recent ychromomal bottlenecks of a single lineage.
You most sorely need to recognize that there is a total difference between genetic relationships and language contact. An example of a what we mean by a genetic relationship in linguistics - you and your sibling share a genetic relationship, you share the same single ancestral origin, but through evolution came to be different entities. For language contact, imagine you and your sibling engaged with each other extensively, and thus came to influence each other, and because of this became more similar in many aspects (words and phrases, ideas, dress). These two discrete things can be with ease distinguished from one another with no reasonable chance of mixing the two in any case. It is very unreasonable to suggest that linguists cannot distinguish extensive contact from genetic relationships. Humans behave in a specific set of universal ways, and for it to be otherwise is an extreme case. This can be applied to linguistics.
I also think you place too much on the lexical basis of a genetic relationship of south and east Cushitic, when we are speaking about morphology. Lexical and morphological influences (when particularly heavy) are quite clear when south Cushitic -> east Cushitic (and the reverse) is concerned. There are east Cushitic languages showing south Cushitic influences, especially in the Samic group including Gabbra and Rendille, and with some extant south Cushitic languages showing east Cushitic influences. Boon also shows clear south Cushitic influences. I think Oromoid language do as well.
Beyond the lexical stuff, if east or south Cushitic had influenced each other to the degree the entire phonological and morphological aspects of their languages cannot be isolated or the direction of influence cannot be ascertained, what we are saying is that
east/south Cushitic are essentially mixed languages - a very, very rare occurrence in history. While not being similar at all - I am bot against the possibility of very early dialectal leveling between embryonic south Cushitic and east Cushitic. But post divergence and collapse of mutual intelligibility and then a leveling, and we are close to be suggesting proto-east and proto-south Cushitic were pidgins/creoles, which the evidence by the longest stretch does not support. You are also extremely overblowing the actual degree of contact between east and south Cushitic.
For a few decades now, linguists have shown they can competently tell apart south Cushitic vs east Cushitic influences quite easily in Nilotic and Bantu languages. One of the most interesting byproduct of this research was the that an entire extinct and unrecorded east Cushitic language is postulated to have existed somewhere in lake Turkana within the borders of south Sudan (named Baz), it's entire existence hinging on it's influences on various Nilotic languages.
The DNA side of things, the early pastoral samples were unlikely to be South Cushitic, and even if they were Cushitic, the fact remains that they have very heavy to predominate non-Cushitic admixture. They belonged to an unknown, and isolated population. Their autosomals were very weird and unexpected, with a ydna that was also very out of place. And most critically,
they didn't share recent ancestry with the south Cushitic samples. They were clearly a distinct, isolated population that
didn't even contribute to the south Cushitic samples.
On the Eurasian %, while I don't agree, all I predict is that the proto-Afroasiatic pop most likely didn't have the same amount of Eurasian ancestry that it's daughter neolithic populations did. I predict an increase in Eurasian ancestry to be one of the critical differences between the proto-Afroasiatic pop and the many neolithic post-breakup Afroasiatic groups. For some groups, additional ancient East African/ANA occurs, as it probably did with early Cushitic, Chadic, and Egyptian groups.
I think also it is very suggestive that
proto-Afroasiatic has a strong possibility that it had a significant degree of langauge contact with the other major African langauge families, and not with the extant language families of Eurasia.
Omotic is 100% Afroasiatic. Anyone who doubts this pretty much is hiding behind racial apologia. It has the largest time depth
only because of its genetic relationship with the other languages. Their E-M34 subclade hasn't been tested yet, but it is probably very old and split very early on from Semitic and later Afroasiatic E-M34.
I prefer we focus on the actual evidence instead of our emotions here. Omotic is Afroasiatic, and it is here to stay.