AlphaProteo can generate new protein binders for diverse target proteins, including VEGF-A, which is associated with cancer and diabetes complications. Notably, this is the first time an AI tool has successfully designed a protein binder for VEGF-A.

The system’s performance is particularly impressive, achieving higher experimental success rates and binding affinities that are up to 300 times better than existing methods across seven target proteins tested:

Trained on vast amounts of protein data from the Protein Data Bank and over 100 million predicted structures from AlphaFold, AlphaProteo has learned the intricacies of molecular binding. Given the structure of a target molecule and preferred binding locations, the system generates a candidate protein designed to bind at those specific sites.

To validate AlphaProteo’s capabilities, the team designed binders for a diverse range of target proteins, including viral proteins involved in infection and proteins associated with cancer, inflammation, and autoimmune diseases. The results were promising, with high binding success rates and best-in-class binding strengths observed across the board.

For instance, when targeting the viral protein BHRF1, 88% of AlphaProteo’s candidate molecules bound successfully in wet lab testing. On average, AlphaProteo binders exhibited 10 times stronger binding than the best existing design methods across the targets tested.

The system’s performance suggests it could significantly reduce the time required for initial experiments involving protein binders across a wide range of applications. However, the team acknowledges that AlphaProteo has limitations, as it was unable to design successful binders against TNFɑ (a protein associated with autoimmune diseases like rheumatoid arthritis.)

To ensure responsible development, Google DeepMind is collaborating with external experts to inform their phased approach to sharing this work and contributing to community efforts in developing best practices—including the NTI’s new AI Bio Forum.

As the technology evolves, the team plans to work with the scientific community to leverage AlphaProteo on impactful biology problems and understand its limitations. They are also exploring drug design applications at Isomorphic Labs.

While AlphaProteo represents a significant step forward in protein design, achieving strong binding is typically just the first step in designing proteins for practical applications. There remain many bioengineering challenges to overcome in the research and development process.

Nevertheless, Google DeepMind’s advancement holds tremendous potential for accelerating progress across a broad spectrum of research, including drug development, cell and tissue imaging, disease understanding and diagnosis, and even crop resistance to pests.

You can find the full AlphaProteo whitepaper here (PDF)

See also: Paige and Microsoft unveil next-gen AI models for cancer diagnosis

Want to learn more about AI and big data from industry leaders? Check out AI & Big Data Expo taking place in Amsterdam, California, and London. The comprehensive event is co-located with other leading events including Intelligent Automation Conference, BlockX, Digital Transformation Week, and Cyber Security & Cloud Expo.

Explore other upcoming enterprise technology events and webinars powered by TechForge here.

The post AlphaProteo: Google DeepMind unveils protein design system appeared first on AI News.

LabGenius is currently focused on discovering new treatments for cancer and inflammatory diseases. The firm combines AI, lab automation, and synthetic biology for its potentially life-saving work.

Until now, the company has been using traditional GPUs for its workloads. LabGenius reports that switching to Graphcore’s IPUs in cloud instances from Cirrascale Cloud Services enabled its training of models to be reduced from one month to around two weeks.

“Previously we used GPUs and it took us about a month to have a functioning model of all the proteins that are out there,” said Dr Katya Putintseva, a Machine Learning Advisor to LabGenius.

“With Graphcore, we reduced the turnaround time to about two weeks, so we can experiment much more rapidly and we can see the results quicker.”

Specifically, LabGenius is using IPUs from Bristol, UK-based Graphcore to train a BERT Transformer model on a large data set of known proteins to predict masked amino acids. This, the company says, enables the model to effectively learn the basic biophysics of proteins.

“[The system] is looking across different features we could change about the molecule — from point mutations of simpler constructs to the overall composition and topology of multi-module proteins,” explained Tom Ashworth, Head of Technology at LabGenius.

“It’s making suggestions about what to design next… to learn about a change in the input and how that maps to a change in the output.”

One in two people now develop cancer in their lifetime. Current treatments often cause much suffering themselves and, while survival rates for most forms are increasing, only around 50 percent survive for ten years or more.

AI will help to find new cancer treatments that cause less suffering and greatly increase the odds of long-term survivability. However, while discovering new cancer treatments is the current focus of LabGenius, the company notes how the principles can be applied more widely to find new treatments for other horrible diseases that plague mankind.

“Graphcore has changed what we’re able to do, accelerating our model training time from weeks to days,” adds Ashworth.

“For our data scientists, that’s really transformative. They can move much more at the speed they think.”

(Photo by National Cancer Institute on Unsplash)

Want to learn more about AI and big data from industry leaders? Check out AI & Big Data Expo taking place in Amsterdam, California, and London.

Explore other upcoming enterprise technology events and webinars powered by TechForge here.

The post LabGenius uses Graphcore’s IPUs to speed up drug discovery appeared first on AI News.

The candidate was identified by scientists who used BenevolentAI’s target ID tools and machine learning models to identify and experimentally validate a novel biological target. Impressively, the achievement was made without any prior reference in published alliteration or patents linking the gene to ulcerative colitis.

Anne Phelan, Chief Scientific Officer at BenevolentAI, said:

“Ulcerative colitis is a chronic, lifelong disease that affects 0.2% of the US population alone and 1.6 million patients in the seven major markets, yet it is poorly served by the standard of care therapies.

Our novel preclinical candidate addresses the high unmet need for an oral, safe and efficacious therapy and has demonstrated improved safety and tolerability profile compared with other leading IBD treatments.

We are actively using patient-derived molecular descriptors to target patient subgroups that will optimise trial design and further increase our probability of success.”

Following the identification of the candidate, BenevolentAI’s molecular design capabilities were used to generate a potential oral and peripherally-restricted candidate drug. The preclinical candidate has been experimentally validated in ex-vivo ulcerative colitis colon samples from patients who didn’t respond to current treatments.

Joanna Shields, Chief Executive Officer of BenevolentAI, commented: “Nominating a drug candidate for a novel ulcerative colitis target identified by our AI-drug discovery platform represents a milestone for BenevolentAI but, more importantly, advances a new potential treatment for this debilitating disease.” 

BenevolentAI plans to advance the asset into clinical trials in early 2023.

(Photo by National Cancer Institute on Unsplash)

Find out more about Digital Transformation Week North America, taking place on 9-10 November 2021, a virtual event and conference exploring advanced DTX strategies for a ‘digital everything’ world.

The post BenevolentAI’s drug discovery platform identifies novel target for ulcerative colitis appeared first on AI News.