Scientists Create An Artificial Neural Network From DNA

Researchers at Caltech have built an AI network using only organic matter. This is the first step in creating a network that could diagnose diseases, think for itself, make decisions and even forge its memories.

At the moment, the study published in the journal Nature by the researchers at Caltech shows that their invention can categorize handwritten numbers. Check out the video posted on the Caltech YouTube channel. There, the artificial neural network classifies the information it receives.

One of the researchers of the study, Lulu Qian, who is an assistant professor of bioengineering at Caltech, explains their work:

“Humans each have over 80 billion neurons in the brain, with which they make highly sophisticated decisions. Smaller animals such as roundworms can make simpler decisions using just a few hundred neurons. In this work, we have designed and created biochemical circuits that function like a small network of neurons to classify molecular information substantially more complex than previously possible.”

It’s a ‘Smart Soup’

The artificial neural network was developed in a test tube and it’s made from synthetic DNA. It basically looks like a soup, only that it’s smart. Scientists used DNA to construct it because single strands of DNA are made from only four molecules – A, T, C and G and their reactions are easy to predict.

The team showed that the AI can be inserted into synthetic biomolecular circuits by using a challenge: molecular handwriting. Molecular handwriting doesn’t have the shape of actual writing. It uses molecular numbers, each number having 20 unique DNA strands, each strand being selected from 100 molecules that represent pixels in a 10 by 10 pattern, all mixed into a test tube.

The ‘smart soup’ can identify the molecular number from nine digits – from 1 to 9.

The neural network first had to separate two digits, and it identified all 36 handwritten numbers through an approach called “winner takes all,” and a type of DNA molecule named “the annihilator.” Kevin Cherry, graduate student and the first author of the study explains the test:

“The annihilator forms a complex with one molecule from one competitor and one molecule from a different competitor and reacts to form inert, unreactive species. The annihilator quickly eats up all of the competitor molecules until only a single competitor species remains. The winning competitor is then restored to a high concentration and produces a fluorescent signal indicating the networks’ decision.”

The next test made the neural network sort between the nine digits it saw and it was successful. Qian said that the potential of artificial intelligence in molecular machines has a great potential:

“Similar to how electronic computers and smartphones have made humans more capable than a hundred years ago, artificial molecular machines could make all things made of molecules, perhaps including even paint and bandages, more capable and more responsive to the environment in the hundred years to come.”

Doris’s passion for writing started to take shape in college where she was editor-in-chief of the college newspaper. Even though she ended up working in IT for more than 7 years, she’s now back to what he always enjoyed doing. With a true passion for technology, Doris mostly covers tech-related topics.