BioMap and MBZUAI team up on joint biocomputing lab to promote sustainable development and human health in the Middle East with large-scale life science models

2023-03-14

Monday, March 14, 2023, Masdar City, Abu Dhabi, UAE: Through a signing ceremony, BioMap, a leading AI life science platform, announced its strategic cooperation with Mohamed bin Zayed University of Artificial Intelligence (MBZUAI), a world-leading artificial intelligence research institution in establishing a joint laboratory. The laboratory will be the first biocomputing innovation research laboratory in the Middle East, and will address the needs of life science scenarios in the region.

Using BioMaps world leading cross-modal bio-computing model, "xTrimo (Cross-model Transformer Representation of Interactome and Multi-Omics) ,the two parties aim to explore new technologies to advance large-scale life science models in protein generation, protein structure prediction, cell function prediction, and other life science tasks. The two parties will focus on driving breakthroughs in AI generated proteins (AIGP) by prioritizing the Middle East's extensive needs in medical health, drug design, energy, and environmental protection.

The signing ceremony was attended by Jiarun Qu, Vice President of Strategic Development of BioMap, and representatives from Mohamed bin Zayed University (MBZUAI). During the ceremony the two organizations introduced their collaboration roadmap, and BioMap shared the technical progress made with its xTrimo model system and its strategic development plan for the Middle East.

Recent improvements in big data collection and computing power have led to breakthroughs in the creation of large-scale pre-training models. Using up to thousands of billion parameters, these models have become especially adept at understanding language and having conversations that seem human-like. ChatGPT in particular has played a leading role in improving model and generation capabilities, making natural language models more easily accessible to the public.

Founded in 2020, BioMap specializes in the research and development of large-scale life science models. Similar in concept to large-scale pre-training models, BioMaps models are not focused on human language processing, but on proteins, the natural language of life. 

As the worlds largest cross-modal life science model, xTrimo models how proteins, cells, and complex biological systems work together, allowing downstream models to learn the operating laws of biological systems and analyze how proteins interact with each other and with cells in the human body. xTrimo is capable of predicting and generating new proteins that can perform targeted interventions and treatments on cells and the human body to address specific needs.

xTrimo has achieved state-of-the-art (SOTA) performance in various areas such as protein structure prediction, antibody sequence generation, and cell characterization. Additionally, it has made breakthroughs in predicting cell function and designing new drugs from scratch.

Highlights of Biomaps xTrimo model:

  •   Achieved an industry-leading AUC in protein design and protein structure prediction accuracy. Its accuracy in predicting antigen-antibody binding far exceeds the industry average, with speeds 100 times faster than the SOTA method.
  • Powerful cell characterization capabilities, along with the industry's most accurate model for predicting synergistic drug effects.
  • Accurate first-time modeling of immune cells and provides precise predictions of functional changes in response to protein perturbation.
  • Extensive evaluation and optimization capability that covers antibody functions, druggability, and more, which has led to the development of a number of self-developed AI drug pipelines within a short time frame.

As announced in the signing ceremony, BioMaps joint laboratory will be in cooperation with the Mohamed bin Zayed University of Artificial Intelligence, a world-class institution focused on artificial intelligence. The collaboration aims to apply AI protein generation capabilities to large-scale life science models to help promote sustainable development and improve human health in the Middle East.

The joint laboratory announced its research would be carried out in two distinct directions: the de novo design of oil degradation enzymes, and mining for potential drug targets for the treatment of aging-related and rare diseases.

In recent years, scientists have been exploring the use of enzymes to efficiently degrade contaminations caused by oil.  For example, alkanes are the major components of crude oils and are commonly found in oil-contaminated environments. Although there are some kinds of alkane oxygenase extracted from microbes, those enzymes usually require a higher temperature to function with ideal catalytic activity, which is actually not the case of marine oil spill, oil and gas storage, and transportation.

Faced with the unmet need, BioMap plans to utilize its AIGP platform to denovo design a new protein, with 10 times improvement of alkane oxygenase catalytic function under lower temperature and even without the help of coenzyme. By using similar methodology, they could enhance the performance and broaden the applications of oil degradation enzymes and oil recovery enzymes by optimizing their catalytic efficiency under different conditions.

Besides cooperating on energy sustainability, the two parties will also focus on scientific research on aging-related diseases, a major public health challenge around the world. In recent years, with the extension of human life expectancy and the increasingly aging population, research on drug targets for aging-related diseases has become one of the hotspots in the field of life sciences.

The two parties will collaborate using large-scale AI models and multi-omics pre-training to accelerate the discovery of drug targets for aging-related diseases, enabling the development of personalized treatments.

BioMap is a disruptive super-scale AI model-powered life science platform. Co-founded by Baidu CEO Robin Li, BioMap aims to leverage cutting-edge AI and biotech to tackle the grand challenge of decoding life and solving the most valuable problems in life science industry, such as target discovery, denovo drug design and enzyme optimization.

With BioMap's AIGP (AI generated proteins) platform, scientists are capable of modeling life from protein to system level, making dialogues with the system via denovo generated proteins, and realizing specific biology functions.