Feed forwards

Apparently, the dynamical system involving only feed forward neural networks (not artificial) converge to a single point. Feedback is required in order to create complex effects, such as convergence to a complex attractors, etc. The structure of genetic networks is similar, and E. coli's regulatory network is largely composed of feed forward networks and very little feedback (except self-loops). Is it possible that for a single input pattern, the cell is adapted to respond in the same way every time, somewhat like a neural network learning an input pattern? 

Is it possible that the genetic network "sets up" the community of proteins in the cell. The community of proteins determine the dynamics of the cell. In this is the case, then the genetic network is a system that takes input from the environment and produces a dynamical system, or  machine, as the output. The dynamical system is designed to survive in the particular input environment. Here is an somewhat odd example (completely hypothetical) :

Lets consider a Venus fly trap or some similar plant...

input: bug lands inside the plant's mouth.

processing: some transcription factor triggered

output: production of several insect digestion enzymes

The output proteins may not just be for digestion itself. They might together form a small machine, such as an oscillator that is responsible for closing the mouth, or some other mechanics that is responsible for carrying out the whole digestion process. By identifying the machine that is the "output" it may be possible to learn behavior pattern is needed to cope to particular environmental signals.