Leveraging the supercomputer, the researchers trained the model on as many as 421,000 human protein pairs.
A supercomputer has helped to create an artificial intelligence model that deciphers the “language of proteins.”Scientists at the University of Glasgow used the Tursa supercomputer at the UK’s DiRAC High Performance Supercomputer facility.
This advanced machine is typically reserved for cosmic research, but it was used for medical purposes in this study. Tursa led to the creation of the protein language model called PLM-Interact. Interestingly, the model predicts protein interactions and even forecasts which mutations will disrupt the communication between these vital molecules.“It’s great to think that DiRAC, which was developed to help scientists understand the laws of nature from the smallest subatomic particles to the largest scales in the Universe, has helped us build this new model to explore the inner space of protein interactions instead,” said Dr Ke Yuan, one of the paper’s corresponding authors. Early tests show PLM-Interact outperforms existing models, including advanced tools like Google DeepMind’s AlphaFold3, in predicting protein interactions.In the field of medical science, the model could offer a better way to understand the development of diseases like cancer and viral infections.Tursa is an Extreme Scaling GPU-based DiRAC system. Credit: EPCCExtensively trained model Proteins are regarded as the workhorses of life, fundamental to cellular structure and all biological processes. These interactions are called protein-to-protein interactions , and any disruption in them often signals the start of diseases like cancer and genetic disorders.Furthermore, viruses exploit these interactions by hijacking host proteins to replicate during infection.Deeper insight into PPIs is essential for advancing the development of new treatments and vaccines. PLM-Interact offers a powerful new lens into these critical mechanisms.The key to PLM-Interact’s success lies in its training. Leveraging the supercomputer, the researchers trained the model on as many as 421,000 human protein pairs. Particularly, Tursa’s computing power was key to the rapid fine-tuning of the 650-million-parameter model. It enabled the model to learn the complex language of protein communication.Predicts protein interactions with accuracy The model predicts protein interactions with 16% to 28% greater accuracy than other cutting-edge AI protein models. PLM-Interact successfully predicted five key protein interactions driving essential biological functions like RNA polymerization and protein transportation.In comparison, other protein AI tools like AlphaFold3 could only manage to predict one of the five protein-to-protein interactions.Moreover, researchers demonstrated PLM-Interact’s ability to identify how mutations influence protein interactions, leading to genetic diseases or inhibiting important PPIs in cancers.When further trained on human and virus protein interactions, PLM-Interact again outperformed existing models in predicting how viruses interact with host proteins. This opens exciting avenues for understanding virus emergence, assessing pandemic potential, and identifying new drug targets to combat viral infections. “The urgency to understand virus-host interactions during the COVID-19 pandemic is a good illustration of why a tool like PLM-interact could be invaluable in the future,” said Prof David L Robertson, corresponding author.Currently, identifying protein interactions experimentally is both costly and time-consuming. PLM-Interact offers a rapid, accurate, and scalable solution, providing scientists with vital new insights into disease and infections.The research team is now looking forward to exploring the full potential of PLM-Interact across a wide range of applications. Eventually, further development could pave the way for an accelerated development of novel therapies, vaccines, and a deeper understanding of life’s fundamental processes.The findings were published in the journal Nature Communications.
Cancer Health PLM-Interact Protein Interactions Protein Language Model Protein-To-Protein Interactions Tursa Supercomputer Viral Infections
United States Latest News, United States Headlines
Similar News:You can also read news stories similar to this one that we have collected from other news sources.
Body Language Expert Claims Princess Kate Tried to 'Ward Off Any Unwanted Attention' During One Awkward Moment With Sarah FergusonAmy Mackelden is the weekend editor at Marie Claire, where she covers celebrity and royal family news. She was the weekend editor at Harper’s BAZAAR for three years, where she covered breaking celebrity and entertainment news, royal stories, fashion, beauty, and politics.
Read more »
View from the Top: Huntsville isn’t just smart - this is why it’s also coolLeveraging cultural heritage is a key ingredient in the playbook of some of America’s most prosperous cities.
Read more »
Supercomputer Predicts 2025–26 Premier League Table As Arsenal Punish Liverpool SlumpManchester United’s strong run of form hasn’t yet translated to the predicted standings.
Read more »
Chicago mayor lashes out at reporter over 'illegal alien' term: 'Racist, nasty language'Fox News Channel offers its audiences in-depth news reporting, along with opinion and analysis encompassing the principles of free people, free markets and diversity of thought, as an alternative to the left-of-center offerings of the news marketplace.
Read more »
AI and quantum tech to drive $1B drug discovery effort in BahrainBahrain’s sovereign wealth fund Mumtalakat inks $1-billion deal with US-based SandboxAQ to accelerate drug discovery, leveraging AI, quantum models, and local health data to develop new therapies.
Read more »
Apache Junction dual language program continues to growNick Ciletti is a 10-time Emmy award-winning journalist who joined ABC15 in July 2014. Currently, you can catch Nick every weekday anchoring ABC15 Mornings from 4:30 a.m. to 7 a.m.
Read more »