New Breakthrough in AI Drug Design | Fapon Digiwiser Releases AI Protein Dynamic Conformation Model DeepConformer

On June 20, Fapon Digiwiser, a subsidiary of Fapon, appeared at the 2024 Conference on Artificial Intelligence and Biomedical Ecology (AIBC), and released the AI protein dynamic conformation innovation model “DeepConformer”, which brought about a major breakthrough in the application of AI in the field of drug design.

At the exhibition, Fapon Digiwiser also demonstrated three innovative platforms, including AI protein engineering platform based on DeepConformer model, AI antibody screening and optimization platform and RNA design AI platform, which fully demonstrated the commercial application prospects of its technology in a wide range of fields such as Biopharma, synthetic biology and immunoassay.

Video for DeepConformer

DeepConformer: a New Approach to AI Protein Dynamic Conformation Prediction

Professor Ma Buyong, founder and chief scientist of Fapon Digiwiser, put forward the conformational selection theory of biomolecular recognition. Conformational selection theory reveals that proteins will selectively adopt specific conformations when performing their functions, so that they can interact with other molecules and complete specific biological functions, which is an important basis from structure-based drug design to dynamic conformation-based drug design. The DeepConformer model released this time further realizes the breakthrough of AI protein dynamic conformation generation, and significantly improves the efficiency of AI assisted drug development.

The most breakthrough of DeepConformer is to use diffusion model as a generative AI framework to directly predict a variety of dynamic conformations and dynamic correlations of protein three-dimensional structures from protein sequences. This prediction method can train protein conformation data independent of molecular dynamics simulation only through the public protein structure database, and the results are similar to those of molecular dynamics simulation

Professor Ma Buyong, Founder and Chief Scientific Officer of Fapon Digiwiser (right)

Efficient Sampling: Refreshing the New Speed of AI Protein Dynamic Conformation

For a long time, the development of original research drugs under the traditional drug design mode has the problems of long development cycle, low success rate and high risk. In recent years, artificial intelligence has been widely used in the drug development process, and has been solving the “pain point” behind drug design.

DeepConformer is a generative AI algorithm model based on protein amino acid sequence to generate the corresponding dynamic three-dimensional structure, which integrates three highlights of fast, efficient, powerful and widely used. The model is trained on the basis of a public protein structure database, a AI assisted protein dynamic conformation technology is adopted, and the time for sampling one structure of the protein with the length of 180 amino acids is only about 4 seconds, so that the conformation sampling efficiency is greatly improved, and the calculation cost is remarkably reduced.

In addition, by identifying and utilizing the different conformational States of proteins, DeepConformer can capture the functional dynamics of proteins more comprehensively, thus providing more targets for drug design, helping to design more specific drug molecules, reducing drug side effects and improving drug efficacy.

At the same time, DeepConformer can be generalized in different protein molecular systems, and can be directly applied to protein design process based on protein structure-activity relationship and prediction and optimization of protein physical and chemical properties.

Professor Ma Buyong said, “This prediction model from sequence to dynamic conformation, you can understand that through AI tools, input text can generate video, the efficiency has been greatly improved.”

Customers carefully read the business introduction of Fapon Digiwiser

Reducing Cost and Increasing Efficiency: Three Technology Platforms Help Drug Design “Fast and Accurate”

The three technology platforms of Fapon Digiwiser have brought new methods to reduce costs and increase efficiency for AI to assist drug design and accelerate drug research and development.

AI protein engineering platform, with dynamic conformation as its core, innovatively combines AI protein dynamic conformation prediction with molecular simulation to form a complementary dual-engine model, and solves the problem of “accurate, not fast enough” by relying solely on molecular simul ation through more accurate algorithms and models. At the same time, it gives full play to the advantages of fast dynamic conformation of AI protein and exploring the conformational space that is difficult to explore by molecular simulation, and achieves the technical goal of “fast and accurate” AI assisted protein prediction.

AI antibody screening optimization platform can quickly screen out high-potential drug candidate molecules through high-throughput, improve the success rate of new drug development, shorten the drug development cycle, and reduce the risk and cost of research and development. By identifying and combining different conformations of multiple targets, drugs with more comprehensive efficacy can also be developed.

RNA as a target compared to the traditional protein targets has its special advantages, the number of its targets is more, and it can make the protein targets which can not be made into medicine, so this kind of drug molecules has attracted wide attention in the industry. The RNA design AI platform of Fapon Digiwiser provides multi-dimensional reference of structure and quantitative indicators for rational drug design through aptamer Al screening and optimization, aptamer + protein binding mechanism analysis, mRNA and circRNA sequence generation and delivery optimization, and structural changes of RNA binding pocket

Fapon Digiwiser AI Protein Dynamic Conformation Prediction and Other Technology Platforms Attract Attention

Industrialization can be expected: Biopharma, synthetic biology, immune detection and other fields are widely used

“This design technology, just like we have created high-quality ‘tires’, these ‘tires’ can be installed on different ‘cars’, significantly improving the speed and performance of ‘cars’,” Professor Ma Buyong admitted.

With solid molecular dynamics simulation and analysis capabilities, the above core technologies of Fapon Digiwiser can be widely used in biological drug development, antibody drug and polypeptide drug development, protease optimization, nucleic acid drug development and other scenarios. On the one hand, it can provide new drug research and development support for large Biopharma companies and biotechnology companies, covering the whole process from target discovery to drug optimization, and provide protein dynamic conformation prediction, optimization and design services for research institutes and enterprises in the field of protein engineering. It can also provide high-specificity and high-affinity diagnostic and therapeutic schemes for aptamers and antibody design of targeted proteins. On the other hand, Fapon Digiwiser will also sell and license the software platform for protein dynamic conformation generation, provide annual subscription and technical support services, and support high-quality investors to create products and share results.

As an innovation-driven life science organization, Fapon has always been committed to the leading underlying technology platform for diagnosis and treatment, combined with the frontier exploration of artificial intelligence in the field of diagnosis and treatment, constantly expanding the boundaries of intelligent diagnosis and treatment in Diagnostics, Biopharma and AI, and jointly promoting the sustainable development of the big health industry with global partners.

AIBC's sharing is full of good materials