The DeepMind research lab has created what is probably its greatest achievement to date: the most complete map of human proteins. Using artificial intelligence, they have been able to predict with great accuracy the protein structure of most of the proteins that make up the human body and that of 20 other organisms. They now plan to make this information freely available to the scientific community.

Proteins are long, complex molecules whose purpose in the body is determined by their structure and how they fold. Their complex and irregular shapes are difficult to analyse and often require long and tedious experiments to find their final form. Why? In order to carry out research into how the body works or to design new drugs.

DeepMind, on the other hand, has chosen a different path: predicting protein structure. Based on all the data that artificial intelligence collects, it is able to predict what the structure and shape of a protein is without having to perform long and complicated experiments. Comparing some of its predictions with proteins that have already been analysed, we can see how the similarity is striking.

Using its powerful artificial intelligence called AlphaFold, DeepMind has analysed virtually all the proteins in the human body (98% of the human proteome). It has also analysed the proteins of 20 other animals common in scientific research, obtaining around 350,000 protein structures per prediction.

All data will be public and free for anyone to download and analyse or use on their own. According to DeepMind, the idea is to continue making protein structure predictions and publish around 100 million of them.

The importance of knowing the structure of a protein

Why is this relevant? Because of the implications it has for scientific research. Knowing the structure of a protein makes it possible to design more effective drugs or, for example, to create crops that are more resistant to adverse environments (or fluorescent to communicate with the farmer). Indeed, DeepMind’s own analyses of protein structures have already been used in important research such as COVID-19.

That said, it is worth remembering that these are predictions and not verified structures. While they can be extremely helpful in giving researchers a general idea or helping them get a better perspective, they are not always 100% reliable. DeepMind’s AI indicates what percentage it believes its prediction to be accurate, but for some cases it will need to verify these protein structures through traditional experiments.

Related

DeepMind creates ‘transformative’ map of human proteins …
DeepMind is releasing predictions for the structure of some 350,000 proteins across 20 different organisms, including animals like mice and fruit flies, and bacteria like E. coli. (There is some…

DeepMind says it will release the structure of every …
DeepMind says it will release the structure of every protein known to science The company has already used its protein-folding AI, AlphaFold, to generate structures for the human proteome, as well…

DeepMind’s AI uncovers structure of 98.5 per cent of human …
DeepMind has mapped the structure of 98.5 per cent of the 20,000 or so proteins in the human body. For 35.7 per cent of these, the algorithm gave a confidence of over 90 per cent accuracy in…

DeepMind AI cracks 50-year-old problem of protein folding …
DeepMind put AlphaFold through its paces by entering it for a biennial “protein olympics” known as Casp, the Critical Assessment of Protein Structure Prediction. Entrants to the international…

DeepMind releases massive database of 3D protein …
On Thursday, Alphabet-owned AI outfit DeepMind announced it has used its highly accurate deep learning model AlphaFold2 to predict the 3D structure of 350,000 proteins — including nearly every…

DeepMind AI predicts protein structures
DeepMind AI predicts protein structures Programs built by Google’s artificial intelligence company DeepMind have beaten humans playing chess, Go, and some Atari computer games. Biologists and computer scientists say the company has now done the same with the protein-folding puzzle.

Deep Mind Knows how Proteins Fold
Deep Mind Knows how Proteins Fold Posted on December 3, 2020 This week, Deep Mind, a London-based Google company, claims to have solved the number one most consequential problem in computational biochemistry: the protein-folding problem. If true, this could be the start of something big.

AlphaFold: Using AI for scientific discovery
Both of these methods relied on deep neural networks that are trained to predict properties of the protein from its genetic sequence. The properties our networks predict are: (a) the distances between pairs of amino acids and (b) the angles between chemical bonds that connect those amino acids.

DeepMind accurately predicts protein structure, advancing …
DeepMind accurately predicts protein structure, advancing decades-old challenge In a challenge where researchers try to predict the shape of newly identified proteins, an AI system developed by…

LEAVE A REPLY

Please enter your comment!
Please enter your name here