Video Script - Biocomputers

What's a Biocomputer?

How many computers are you using right now? I mean, clearly you're using a computer to watch this video, but how many are you not aware of? There's actually a lot of computers around us. Take a look inside your phones, tablets, and laptops and we'll all know of the CPUs inside their noggins. But what about the computer inside your microwave (to help control the time and temperature of your food), your refrigerator (making sure things are cooled and maintain homeostasis), your car, your TV, your thermostat, ...

Especially, the one rIgHt beHinD yOu...

Naw I'm just kidding.

We'll there's one that I bet you weren't thinking of, your brain! You're brain is technically a kind of computer. To be semantic, let's look at the definition of what it means to be a computer. Ahem (pulls out sheet of paper):

one that computes (https://www.merriam-webster.com/dictionary/computer)

well yeah no shi--

So we need a different definition. Really, the most general definition of something that we can call a computer is something called a Turing Machine. It's something that simply:

• Can read data
• Can write data
• Can do conditional actions when reading that data.

This is very loose definition, but you can catch my drift. So remember 30 seconds ago when I said that a brain is a computer? We'll it fills out all the criteria:

• Can read data (read inputs from the world via our senses)
• Can write data (neurons save information in our memories)
• Can do conditional actions when reading that data (we do actions by the world we perceive around us).

The Why

So why am I bringing any of this up at all? We'll, while computers are very good at many things, they are very limited at doing certain things that humans, or anything with a brain, can do. For one, they don't have consciousness, while we do. I think we can all agree on that.

We'll, unless your this guy.

But one other interesting consideration is the power management of human computers, which we'll denote biocomputers (I know, humans aren't the only ones to have brains), versus classical computers, the ones in your phone. Mainly, biocomputers are very power efficient.

I mean, think about it. If one needs 2000 cal = 8,368,000J per day (86,400 secs) that's about 96.9W! and it's estimated that about 20% (see Figure 5) of the energy we use in a day goes to the brain. So that's about 20W per day needed just for our brains. In contrast, most desktop PC's require over 65 to 150 watts, depending on the performance of your processor.

Furthermore, they are still small form factor, on the order of micrometers, as we'll see in a bit. And another big reason to look into making them is that the "attempts to develop computer electronics that emulate the circuitry of the brain may lead to a better understanding of how the brain itself actually works", as Professor Jorgensen, the director of the Health Physics Graduate Program at Georgetown University, has put it.

So, there's reason to try to make computers out of biological materials (not just make human computers solely). Let's look at a few groups who've taken the undertaking to make some of these things.

Who's Making Them? Into What?

Our first group is a BioEngineer research group from Stanford who developed the concept of a transcriptor. You see, in classical computer logic we use something called a transistor, which is a voltage-controlled switch. This is required to do digital logic for things called gates, which in combination can do things like add, subtract, redirect signals, and even store data.

The transcriptor is a "control signal that regulates the number of RNA polymerase molecules flowing through a separate DNA element". We have a molecule of DNA, and use RNA polymerase as the 'electron' in this case. Essentially, the RNA polymerase will only flow along the transcriptor when the control signal is set.

However, what differs transcriptors form their transistor counterparts is we are allowed to stack control signals. This allows us to, say, make an XOR gate using just one transcriptor and two control signals, stacked to work on the same transcriptor. As a result, we can combine gates and make use for more space. These new gates are called BIL Gates, or Boolean Integrase Logic gates.

No not you!!. Get out of here Bill!

But we can do even more than just do logic! Just like their transistor counterparts, we can actually amplify an input signal using a BIL gate. Essentially, the "switching" action of these transcriptors allow signals that are near the "switching" point to get extremely amplified. As a result, signals can be amplified by keeping the input signal near that switching point.

We also have other groups that have made pieces of our current technology. Here at MIT another group found a way to make computer memory out of transcriptors. Their structure and function are essentially the same as their transistor counterparts.

Interestingly, over their classical counterparts, this biological method helps store more data for cheaper. One reason is that classical computers are base-2, so the wires are only on or off. On the other hand, DNA is base-4 (since we have 4 nucleic acids to choose from) and as such large numbers in base-2 are a smaller number of digits in base 4. To illustrate, to store the number 255 in base-2, that's:

while if we denote $A:=0,C:=1,G:=2,T:=3$ then storing the same number would be:

Notice how much fewer digits are needed to store the same number between both systems!

In Conclusion

We looked at a few reasons why transcriptors can be the future in regards to data storage; however, there are massive limitations that make it likely to not enter your computer any time soon.

For one, while the size of these BIL gates are on the orders of cells, so a few micrometers, current transcriptors are on the order of nanometer technology. They are considerably bigger, and as a result can tend to hold less data when in larger quantities (see Figure c):

As a result, they also will likely not be entering the CPU, GPU, or other processing unit spaces anytime soon.

So while we won't likely be seeing artificially lab-grown brains anytime soon, it's still very impressive to see where the technology has taken us.

Thanks for listening to my TED-Talk!