Shulou(Shulou.com)11/24 Report--

On November 1-3, 2023, a training meeting focusing on frozen electron microscope single particle analysis technology was held in Southern University of Science and Technology, which was jointly organized by Volcano engine, Cryosurgery Center of Southern University of Science and Technology, Shanghai New Life Department and Intel. Participants discussed in depth the broad prospects and applications of frozen electron microscope technology in basic research, clinical research and innovative biological drug development.

At the meeting, many famous experts and scholars in the field of cryoelectron microscopy from Southern University of Science and Technology, Fudan University, Sun Yat-sen University, Beijing Institute of Technology, Hunan normal University, Tsinghua University and other universities shared wonderful academic reports. In the process of computer training, through patient explanation, the training teacher provides direct guidance and new ideas for students to understand frozen electron microscope, expand research ideas and carry out technical research and development.

Gao Qing, general solution architect of volcano engine, delivered a speech on the theme of "new paradigm of volcanic scientific research cloud-assisted scientific research", showing the research and development achievements of volcanic engine in the field of bioinformatics computing and life science in recent years. IT&BT is accelerating its integration, that is, using digitization to empower genetic research, cell engineering, protein structure prediction and other biotechnology fields to achieve innovation, efficiency improvement, engineering and industrialization, and among the multiple paths of digital innovation, the best way to improve efficiency is on the cloud.

Autonomous and controllable big data operating system (Bio-OS)

Volcano engine is a cloud service platform under byte jump, which provides technical support for various applications of Douyin Group internally in the process of digital innovation, and creates the second growth curve with Qianhang Baiye externally. In the life science industry line, Volcano engine launches solutions in four sub-areas: scientific research, genetics, medicine and systems, providing basic resources such as reliable computing, storage, cloud native operating systems, as well as R & D, data and artificial intelligence capabilities.

Overview of Volcano engine Life Science and Pharmaceutical Industry

From the field of bioinformatics gene computing, with the development of technology, the price of modern whole genome sequencing has broken Moore's Law, and the inequality of research computing resources is increasing. At the same time, massive data storage, efficient transmission and data element sharing in the industry have become important challenges.

Based on the common problems of the industry, aiming at the pain points such as poor repeatability of scientific research achievements, non-uniform data standards, lack of industry standardization, difficult data aggregation, weak analytical engineering ability, etc., Volcano engine, together with partners, developed a set of biological big data operating system based on international standards-Bio-OS, which realized the unity of standard specification, resource tools, development environment and technical architecture.

Bio-OS can help biologists, data scientists and IT personnel to quickly build the corresponding analysis workflow, shielding the tedious problems of environment construction, data search and big data computing task scheduling, greatly simplifying the whole research process and improving the efficiency of the analysis process. In addition, the volcano engine also collates and collates all kinds of high-value data sets, territorial workflows and tools around the world, and queries and invokes the suppliers. At the same time, the kernel of the Bio-OS platform is open source, and it is expected that through the open form, we can build the industry ecology and realize the "universal benefit of scientific research resources".

Three core competencies boost AIfor Science

At present, the technological achievements of artificial intelligence-driven scientific research (AI for Science, referred to as AI4S) play a more and more important role in the fields of biomedicine, materials and physics, and how to use large models to empower scientific research and industrialization has become the core issue of industry reform.

Therefore, based on the basic logic of AI4S, the volcano engine begins to explore the "digital twin" in the field of life science, hoping to form a digital physiological human model that can carry out simulation experiments by integrating and training all kinds of multimodal data, such as genes, proteins, cells, phenotypes, literature, medical records and pathology, so as to help the scientific research team enlighten the research direction and improve the efficiency of disease research and pharmaceutical research and development.

After several years of project practice, volcanic engines have gradually accumulated three core competencies in this field:

Core competence one: cloud native scheduling system

Build and make use of the national computing network, and use cloud native containers to schedule different regions and heterogeneous computing power for fast task delivery. For example, Bio-OS cloud service has provided services to many enterprises and scientific research institutions such as agricultural breeding, third-generation sequencing and AI protein design, providing cloud users with a complete set of solutions for data transmission, data storage, to secondary analysis, tertiary analysis and data delivery, helping to make a breakthrough in the genome field.

Core competence II: tool chain system based on large model

Use the large model to train the high-value data in a certain field, and use the vector library to reduce the hallucination problem, help the research team to explore the research ideas, and generate and reproduce the paper code and data set. For example, the volcanic engine helps a laboratory to explore the innovative scientific research ideas of paper retrieval and experimental data by using the volcanic Ark model platform, and uses the knowledge base for effective professional verification. to promote the connection and exploration of cross-domain knowledge in rice planting and scientific research.

Core competence 3: constantly expanding territorial ecological applications based on volcanic scientific research cloud platform

Together with Yuexin technology in the field of frozen electron microscopy, the volcano engine carries out the engineering transformation of multiple computing tasks based on workflow for Relion. In addition, it also works with Intel to implement end-to-end high-energy optimization of Alphafold2 using the Xeon extensible platform as the hardware base to accelerate protein structure prediction services.

Yang Wei, an Intel artificial intelligence architect, presented the actual results of AlphaFold2 based on volcano engine SPR cloud instance G3i at the conference site, and explained in detail the solution based on Intel fourth generation Xeon processor and OpenVINO ™tool suite. Intel works deeply with Volcano engine to accelerate the application of AI for Science life science computing.

END

In the field of frozen electron microscopy and structural biology, the volcano engine is expected to join hands with more academic institutions and technology manufacturers to jointly create frozen electron microscope clouds by making use of basic capabilities such as digitization and cloud computing. Help the scientific research team solve the core problems such as computing queuing and data storage, and finally promote the accelerated maturity and application of engineering and industrialization capabilities in the whole field.

Welcome to subscribe "Shulou Technology Information " to get latest news, interesting things and hot topics in the IT industry, and controls the hottest and latest Internet news, technology news and IT industry trends.