Ah, you're right...hard to read the Erica stuff when it's itsy-bitsy...
You might not want to keep the subchains as separate as that. It may make more sense from the point of interaction between parts to work out some sort of structure that allows more than just a time-base between all parts. For example, a shifting LFO curve in one part could, via a comparator, switch on an action in a second part when the LFO was above (or below) X level. Say, if your VCF cutoff in the first part went above a certain partial frequency via that LFO, the same LFO could start a sequence via the comparator gate, then stop it when the VCF's cutoff dropped. And going further, that triggered sequencer part could, in turn, activate some other timing function (such as ratcheting) on another sequenced part. And so on, ad infinitum. Best way to proceed, from my experience, is to start by mapping the basic action of a part, then identifying points in that part's signal and/or control chains where the insertion of something that 'reads' the activity could be used to interact with another. And at the same time, identifying points of 'action change' in the affected parts, and figuring out how you want the changes to occur, etc. Usually, these sorts of system require a great deal of control density, which you're definitely getting toward in this build, but it might make sense to spread out more for the ease of programmability and interaction with the system as a whole. Since you like the Erica stuff, have a look at some of their cases...the deep 2 x 126hp will allow you to stretch out more, add a few more bits, and provide more than adequate power while avoiding any depth conflicts.