Demis Hassabis, CEO of Alphabet, Google DeepMind Research Group, at Google’s Future of Go Summit in China on May 23, 2017.
LONDON – Alphabet-owned DeepMind has developed artificial intelligence software that can accurately predict the structure into which proteins will fold in a matter of days, solving a 50-year-old ‘big challenge’ that could pave the way for better understanding of diseases and drug discovery.
Each living cell contains thousands of different proteins that keep it alive and healthy. Predicting the shape a protein will fold into is important because it determines its function, and almost all diseases, including cancer and dementia, are related to how proteins work.
“Proteins are the most beautiful and magnificent structures and the ability to predict exactly how they fold up is really very, very difficult and has occupied many people for many years,” said Professor Dame Janet Thornton of the European Institute of Bioinformatics.
The “AlphaFold” AI system from the British research laboratory DeepMind was entered in a competition organized by a group called CASP (Critical Assessment for Structure Prediction). It is a community experimentation organization whose mission is to speed up solutions to a problem: how to calculate the 3D structure of protein molecules.
The CASP, which has monitored progress in the field for 25 years, compares competition submissions with an “experimental gold standard.” On Monday, he said that DeepMind’s AlphaFold System has reached unprecedented levels of precision in predicting protein structure.
“DeepMind has taken a leap forward,” said Professor John Moult, who is the president of CASP, during a press call ahead of the announcement. “A great 50-year challenge in computing has been solved to a large extent.”
Moult added that there are “major impacts later on in drug design” and in the new emerging field of protein design.
With around 1,000 employees and virtually no income, DeepMind has become an expensive business for Alphabet (Google’s parent company). However, he became one of the leaders in the global AI race with Facebook AI Research, Microsoft, and OpenAI.
The breakthrough was hailed by Google Managing Director Sundar Pichai on Twitter.
Demis Hassabis, co-founder and CEO of DeepMind, said on the call: “DeepMind’s ultimate vision has always been to create general AI and then use it to help us better understand the world that surrounds us. surrounds by dramatically accelerating the pace of scientific discovery. ”
The company, which Google bought for $ 600 million in 2014, is best known for creating AI systems that can play games like Space Invaders and the old Chinese board game Go. However, it still has says he wants to have more scientific impact.
“Games are a great testing ground for effectively developing and testing general algorithms that we hope will someday transfer to real-world areas such as science problems,” Hassabis said. “We believe that AlphaFold is a first point of proof for this thesis. These algorithms are now becoming mature and powerful enough to be applicable to really difficult scientific problems.
DeepMind also entered a CASP protein folding competition in 2018. While his results at the time were impressive, John Jumper, AlphaFold manager at DeepMind, said the team knew he was sort of to produce something with “very strong biological relevance or to be competitive with experience.”
However, this year’s competition was not straightforward, and Jumper said DeepMind went three months without making any progress. “We sit there and worry that we’ve run out of data?” he said.
Even as the competition deadline approached, Jumper and his team were still concerned that they had made mistakes. “There can always be an error seeping into machine learning systems,” he said.
But their efforts seem to have paid off. “We really think we have built a system that provides correct and actionable information to experimental biologists,” he said. “The reason you have structure is to understand something about the natural world and ask more questions. We think we’ve built a system that will really help people do that.”
