Protein Structure Analysis and Molecular Interaction Prediction
Research Achievements Validated by Publication in Nature
Artificial intelligence (AI) research company Google DeepMind has unveiled ‘AlphaFold3,’ an AI model designed to develop drugs and assist in more effective disease treatment.
Demis Hassabis, CEO of Google DeepMind, said in an interview with Bloomberg TV on the 8th (local time), "Google DeepMind has been expanding the horizons of biological research using AI, and AlphaFold3 represents the latest stage of these efforts," adding, "We expect that within the next few years, drugs initially designed by AI will be administered to patients."
AlphaFold3 is the latest model in the AlphaFold series, which identifies protein structures within the human body. Cells involve the complex interaction of billions of molecules such as proteins and genes (DNA), and AlphaFold3 can predict the structures and interactions of nearly all biomolecules fundamental to life. DeepMind explained that predictions regarding interactions between proteins and other molecules have improved in accuracy by more than 50% compared to previous models.
AlphaFold3 is also reported to be capable of modeling chemical modifications of molecules that can disrupt healthy cellular functions and lead to diseases. DeepMind expects AlphaFold3 to contribute to new drug development and disease treatment research in the future. It is known that Isomorphic Labs, a drug development subsidiary of Google’s parent company Alphabet, is currently using AlphaFold3 for drug development.
AlphaFold3 was also published in the international scientific journal Nature. CEO Hassabis said, "We are very pleased to publish a new paper containing these groundbreaking research results in Nature," and evaluated that "the time and cost required to develop treatments that could potentially change our lives will be reduced."
On the same day, Google DeepMind also announced the ‘AlphaFold Server,’ a nonprofit research support platform. Scientists worldwide can use the platform’s features, such as molecular interaction prediction and chemical structure modeling, free of charge.
Meanwhile, as AI technology advances, competition among big tech companies for new drug development is becoming increasingly fierce. NVIDIA, a leader in AI chips, introduced ‘BioniMo,’ a generative AI-based drug development platform that predicts protein structures, in January. Microsoft (MS) also unveiled ‘EvoDiff,’ an AI that generates new proteins, in September last year.
© The Asia Business Daily(www.asiae.co.kr). All rights reserved.
