Okay, here's a simple true/false question: If the Metazoa we see today was the intended outcome of a front-loading scenario, could we make testable predictions from this premise?
In order to be testable predictions they would need to differ from the assumption and differ from known natural mechanisms that you are seeking to replace. So far, you have fulfilled none of these requirements.
No, non-teleological models do not predict that crucial eukaryotic genes will share deep homology with functional but unnecessary (for life) prokaryotic proteins.
Just like non-teleological meteorology does not predict that it has to rain on July 15th, 2016 in Dallas, TX. However, rain that day is consistent with non-teleological meteorology. In the same way, "crucial eukaryotic genes will share deep homology with functional but unnecessary (for life) prokaryotic proteins" is also consistent with non-teleological evolutionary mechanisms.
It is incumbent on you to show that non-teleological mechanisms can not co-opt proteins in subsequent generations.
Read that carefully, then tell me that it's not reasonable, under non-teleological models, for the LUCA to have only a minimal genome and be only a minimal cell.
It's not reasonable under non-teleological models for LUCA to have a minimalist genome. LUCA is a product of many rounds of evolution. LUCA will contain genes that are unnecessary for life by necessary for outcompeting other organisms in a given environment.
What you label unnecessary proteins are beneficial proteins to the organism that carries them. That is why they have been preserved for 3 billion years of evolution. Can non-teleological mechanisms preserve beneficial proteins? Yes. Therefore, it is not a valid testable prediction for FLE.
Can non-teleological mechanisms co-opt beneficial mutations for new purposes through random mutation and natural selection so that they are necessary genes in subsequent generations? Yes, absolutely. This too is not a valid testable prediction for FLE because it does not differentiate FLE from non-teleological processes.
The issue here really isn't whether the non-telic view of life can potentially explain the observation that crucial eukaryotic proteins share deep homology with functional but unnecessary prokaryotic proteins.
Yes, it is. If non-teleological mechanisms can produce the observations then FLE is not supported.