Even if specific proteins are needed for memory creation (which is debatable [1]) it doesn't mean you need to model those proteins to retrieve the stored memories from the structures that they created. You can read data from a hard drive without modeling the CPU or memory bus of the computer that stored the data.
> support cells (glia)
Glia cells are the order of 40-50 microns [2] and can easily be seen with an electron microscope. In fact, they are present in the Lichtman paper [3] linked from the Mind Emulation website.
[1] https://www.ncbi.nlm.nih.gov/pmc/articles/PMC2745628/ [2] https://psych.athabascau.ca/html/Psych402/Biotutorials/4/par... [3] https://www.cell.com/cell/fulltext/S0092-8674(15)00824-7
Anyway, my point is that even if you are able to recreate the structure you would need to replicate the functionality.
A better analogy would be that even if you have the data on the hard drive you would need a special program that would know how to interpret it. In this case the data also contains a time variant exactly when the data applies which is not captured within the structure unless you go to the molecular level.
1. We still don't know what the memory engrams truly are ( https://en.m.wikipedia.org/wiki/Engram_(neuropsychology) ). I once read that aside from relying on the interconnections of the neurons they also rely on specific proteins created during memory creation. They are then vital for memory recollection.
2. We know that the connections between neurons are important but we just realized that the support cells (glia) also affect the firing mecanism: they are not only support but a filter as well ( https://en.m.wikipedia.org/wiki/Glia )
3. Inhibitory interneurons provide a way for synchronous firing of neurons to form a learning experience: https://www.nature.com/articles/s41467-019-13170-w
All in all to replicate the brain functionality we would need to fully replicate the chemical composition of the brain to the lowest level (molecules).
I'm not holding my breath.