A new breakthrough in AI-powered drug design — Fapon Digiwiser launches DeepConformer, an AI model for predicting the dynamic conformations of proteins

On June 20, Fapon Digiwiser, a subsidiary of Fapon, attended the 2024 Artificial Intelligence for BioPharma Conference (AIBC) in Shanghai and launched its innovative AI model DeepConformer which can be used to predict the dynamic conformations of proteins, a major breakthrough in AI-powered drug design.

During the event, Fapon Digiwiser also showcased three AI-powered platforms: the protein engineering platform based on DeepConformer, the antibody screening and optimization platform, and the RNA design platform, demonstrating the extensive commercial applications of its technologies in a wide range of areas such as drug development, synthetic biology, and immunoassay.

DeepConformer: Blazing a trail in using AI to predict the protein’s dynamic conformations

Dr. Buyong MA, Founder and Chief Scientist of Fapon Digiwiser, proposed the conformational selection theory, which is instrumental in deciphering how drug molecules identify and bind to their targets. The theory reveals that proteins would choose specific conformations to interact with other molecules and perform specific biological functions. The theory lays a solid foundation to allow for a shift from a structure-based approach to a conformation-based approach in drug design. DeepConformer launched at this year's AIBC is a model that represents a breakthrough in using AI to predict the dynamic conformations of proteins, which can significantly improve the efficiency of drug development.

DeepConformer is revolutionary because of its ability to directly predict the dynamic conformations and correlations of a protein’s 3D structure based on the protein's sequence by using a diffusion model as a Generative AI framework. Trained with publicly-available protein structural data, this method is capable of generating conformations data similar to those from molecular simulation.

Dr. Buyong Ma (right), Founder and Chief Scientist of Fapon Digiwiser

Efficient sampling: Setting a new record in the speed of AI-powered prediction of protein’s dynamic conformations

Traditional drug design models have long been troubled by issues such as long development cycles, low success rates and high risks. In recent years, AI has been widely used in drug development, and has resolved many pain points in drug design.

DeepConformer adopts a Generative AI algorithm that generates the dynamic 3D structure of a protein based on its amino acid sequence. It is fast, powerful, and has wide applications. Trained with publicly-available protein structural data, this AI-powered tool is able to sample the structure of a protein with 180 amino acids in just 4 seconds, significantly improving the sampling efficiency and reducing the computing cost.

In addition, by identifying and utilizing the different conformations of a protein, DeepConformer is able to comprehensively capture functional dynamics, so as to provide more targets for drug design and help design drug molecules with higher specificity, less side effects and better efficacy.

Besides, DeepConformer can be generalized for systems with different protein molecules, and can be directly used for multiple tasks such as protein design based on structure–activity relations as well as the prediction and optimization of a protein’s physicochemical properties.

Dr. MA said: "You can understand this sequence-to-conformation model as a text-to-video AI tool. It’s going to greatly improve productivity."

A client is carefully reading a leaflet of Fapon Digiwiser

Three platforms to make drug design more efficient and cost-effective

Fapon Digiwiser makes drug design and development more efficient and cost-effective with its three AI platforms.

The AI-powered protein engineering platform based on dynamic conformations innovatively combines the AI-powered prediction of a protein's dynamic conformations with molecular simulation. Conventionally, molecular simulation is accurate but not fast enough. But adding the AI-powered prediction tool not only improves the speed but also probes into the conformational space which can be difficult to reach by molecular simulation. Therefore, this platform is a real breakthrough. It makes the prediction of a protein’s dynamic conformations fast and accurate.

The AI-powered antibody screening and optimization platform can quickly identify high-potential candidate molecules through high-throughput screening, which increases the success rate, shortens the development cycle, and reduces risks and costs. By identifying different conformations of multiple targets, scientists can leverage this platform to develop drugs with better therapeutic effects.

RNAs have their unique advantages over traditional protein targets. They can target a larger amount of targets and make undruggable protein targets druggable. Therefore, RNAs have attracted a lot of attention in the industry. Fapon Digiwiser's AI-powered RNA design platform screens and optimizes nucleic acid aptamers, analyzes the aptamer-protein binding mechanisms, as well as optimizes the sequence generation and delivery of mRNA and circular RNA sequences. These functions, combining the structural changes of RNA binding pockets, provide insights from various aspects, including structures and quantified metrics, for drug design in the real environment.

Many people take interest in Fapon Digiwiser's DeepConformer and other technology platforms

Promising applications in drug development, synthetic biology, immunoassay, etc.

"This design technology is like that we make high-quality tires which can be mounted to different vehicles to significantly improve the speed and performance of vehicles," said Dr. MA.

In combination with Fapon Digiwiser's solid capabilities in simulating and analyzing molecular dynamics, Fapon Digiwiser’s technology platforms can be widely used in the development of biologics, therapeutic antibodies, polypeptide drugs, and nucleic acid drugs, as well as the optimization of proteases. On the one hand, Fapon Digiwiser can provide support in new drug development for big pharmas and biotech companies, all the way from target discovery to drug optimization, provide services for research organizations or companies in the field of protein engineering to predict, optimize, and design the dynamic conformations of proteins, and provide high-specificity and high-affinity therapeutic solutions for those designing nucleic acid aptamers or antibodies targeting proteins. On the other hand, Fapon Digiwiser will sell and out-license its dynamic conformations software, offer annual subscription options and technical support, and partner with quality investors to develop better products and share achievements.

As an innovation-driven life sciences organization, Fapon Group is pushing the boundaries of diagnostics, pharma and AI-driven diagnostics and therapeutics by leveraging its leading technologies and platforms. It is also working with partners around the world to drive the sustainable development of healthcare.

Someone is giving a presentation at AIBC 2024

About Fapon Digiwiser

Fapon Digiwiser is an innovative biotech company committed to integrating artificial intelligence (AI) with molecular simulation to advance the design and regulation of macromolecules by leveraging high-performance computing (HPC) and AI.

As an industry leader in protein dynamic conformations, we push the boundaries of life sciences through DeepConformer and molecular simulation, enabling accurate and efficient prediction of protein dynamic structures. Additionally, we have a competitive edge in capturing the dynamic properties of protein molecules and exploring their active sites and functional mechanisms.