intech.media

Discovering tomorrow's global healthtech trends today

intech.media

Discovering tomorrow's global healthtech trends today

AI Helps Scientists Discover What Proteins Are Hiding

Artificial intelligence has long been revered as an interesting way to process data and information. However, it’s becoming more and more useful in other areas, including science. Scientists have been using Google’s DeepMind to train an AI system to work out a problem that has stumped them for decades: proteins.

Proteins exist in every living thing in the world and are essential for managing chemical processes. They can fold into an infinite number of ways that determine how they’ll function in the body.

 

What’s Google’s DeepMind?

DeepMind is something that Google put together to create safe artificial intelligence systems. The main goal behind the design was to boost science and to create an opportunity for scientific discovery.

The concept behind creating safe AI is that AI can have some pretty amazing benefits but it can also negatively impact the world if not carefully taken care of. Google wanted to be able to create a system that would solve problems and come up with new ideas, but not become biased or unfair. They put together an ethics team that would be focused on keeping the system ethical and safe.

One way to help ensure the entire process is secure, the team anticipates risks, both long and short term, then works out ways to prevent these and manage the risks as needed. They work together with Google, Alan Turing Institute, and OpenAI.

Many groups have used DeepMind to find solutions in Science, but AlphaFold was one of the biggest projects ever. It was a problem that scientists had yet to figure out, despite 50 years of attempting various ideas and theories. After some additional research, they thought an AI system might be able to manage to find the solutions that were eluding them.

Proteins, DeepMind, Google, technology, science, study, AI, DNA, AlphaFold, RNA, deep mind, discovery, AIin science, folding proteins

The Mystery of Proteins

What did scientists need to figure out about proteins? The big question that was haunting them was how protein molecules actually take shape. As the basic part of life, scientists knew that proteins were important and that DNA blueprints worked to tell cells to put amino acids together to create specific types of proteins. What no one knew was how those strings of amino acids folded into complex 3D structures that can be used in the body.

While scientists could see the complex structure, they didn’t understand how to predict the protein folding from the DNA sequence. It was an area that was definitely lacking in their knowledge of the body. This is why a few scientists decided to create AlphaFold, the AI algorithm that would figure out how to predict what type of protein would be made from a string of amino acids. In 1972, Christian Anfinsen determined that you should be able to look at the sequence of amino acids strung along in a protein and determine what shape it would take once folded. He won a Nobel prize for this discovery, but no one was able to figure this out.

Teams of scientists have been working on the problem for the past 50 years, without ever solving it. However, with this final attempt with AI, Google’s DeepMind AlphaFold managed to determine the shape of 66% of the proteins. What had taken scientists years to get close to had been solved in just a few days by artificial intelligence?

To be fair, it wasn’t completely solved, but artificial intelligence was able to figure out most of the ways the proteins fold. This is far more than scientists were able to do in the previous five decades. Thanks to machine learning, scientists could now use AI predictions to reasonably predict how the protein strands will create a 3D structure.

Proteins, DeepMind, Google, technology, science, study, AI, DNA, AlphaFold, RNA, deep mind, discovery, AIin science, folding proteins

How AlphaFold Benefits Science

Knowing how to predict what proteins will look like once folded means we can start to figure out what they do. Proteins are essential in chemical reactions throughout the body, which means many diseases are related to how they work. For example, insulin is a hormone that is produced through the actions of proteins. When we know which protein chains will form into a 3D protein that will affect insulin, it becomes easier to see where things are failing in the body.

A tiny change to any string of amino acids can end up causing some pretty drastic health effects. That’s why scientists are always studying how proteins are involved in many diseases, particularly those where hormones, enzymes, or antibodies are malfunctioning.

Now that we know what the 3D structure of a protein looks like, we can figure out more about a number of diseases. These include dementia, cancer, and other diseases. It’s easier to manage diseases when you know how they’re coming about.

Protein structure is also essential for drug design. Many drugs require the use of proteins, but it’s only possible to use them when you know what they do and how they will react. This is one area that has been studied during the COVID-19 pandemic.

 

What Now?

Through the information provided by AI, we know we can predict many proteins, but that doesn’t mean scientists know everything. In fact, they are still learning how proteins of different types work together. They also need to learn more about the interactions proteins have with RNA and DNA.

Understanding how our bodies work helps scientists come up with better solutions to health issues. While many areas are now understood, proteins have been one of the few parts of the body that still eluded scientists. Now, that’s changed.

It took something man had created to determine what the protein folding process is. Artificial intelligence was the option to help scientists resolve an age-old problem, and it has proven to be the right choice. Where our human minds have failed, the unwavering, rapidly moving functions of a computer have been able to create a solution to a very confusing and complex problem.

〈Related Articles〉

Leave A Reply