First of all - in an attempt to TRY to get to the basics of the OP...
Conductive materials of ANY kind have "X" resistance per unit of length. Resistance in power wires is the enemy which I think we can all agree on.
Copper wire resistance of a given gauge is essentially constant.
Steel/aluminum/tin/lead/cast iron/weld etc of a car's chassis varies widely depending on the car...
BUT - the few tests that have been done indicate the AVERAGE resistance of a car's chassis between typical (-) return paths is somewhere in the range 4-1/0 awg copper wire.
SO -- If your amps require more current than 4 awg wire can supply it might be worth considering a large (-) wire directly to the battery. Larger than 1/0 and it's probably worth doing.
That's really all there is to it.
...any drawbacks of running the power and ground wire together.
None at all.
I wonder though if the EMI would at least partially cancel since you'd have current flowing in equal/opposite directions??
Obvously when you run your RCA's/remote and power, you keep them seperate.
you don't have to seperate those either. Any EMI from the low voltage low freq variances in the power wire are not going to have an audible effect on the signal. Old myth - it just doesn't happen.
The reason it is taught that you shouldn't hook it up directly to the negative terminal for the last connection is because it may spark. In open-cell batteries (not commonly used in car audio, but regular car batteries that 99% of people have), there may be flammable gasses being emitted. ...why take the risk of getting a spark right in the middle of a cloud of flammable gas? This is why they will tell you to connect it to the frame 'as far away from the battery as possible'.
Precisely -- more specifically, hydrogen. Heard of the Hindenburg?