Week 7 Lab: Neuromorphic Circuits

Intracellular Artificial Neural Networks (IANNs)

In this two-day lab, we designed and built our very own IANN using a library of plasmids from the Ron Weiss lab and human embryonic kidney (HEK) 293 cells. IANNs differ from traditional synthetic genetic circuits because IANNs can perform analog computations, rather than being limited to digital computations. IANNs are also universal function approximators—given an adequate number of intracellular artificial neurons, you can use an IANN to achieve any input/output behavior you’d like.

Day 1

I worked with a group of three, and this was the circuit we designed:

Here are the predicted results by the biocompiler:

Day 2

On day 2, we took a trip to the Weiss Lab and Evan started the protocol on the Opentrons machine to begin construction of our neuromorphic circuits. We also looked at some immortalized human cells on a microscope, giving insight to experimental mammalian cell biology.

Results

Our results didn’t quite match the prediction. This is due to the fact that our uploaded circuit violated some of the “laws” governing the system’s design (we learned this later), but the computer still allowed these “illegal” circuits to be generated and simulated.

Gemini AI was consulted for formatting and content organization