In the AI-driven era, human health will “know the difference from the smallest details”
Author：Eve Cole
Update Time：2024-11-22 12:54:02
In the era of artificial intelligence (AI), biological big data, with the support of computing power and algorithms, is pushing the field of human health to a new realm where we can see the subtle changes and the autumn leaves. At the 2024 Pujiang Innovation Forum "Shanghai International Computational Biology Innovation Forum" held today (September 9), focusing on the theme of "Computational Biology Empowers Biomedicine Innovation", the participating guests shared the application of "AI+" in biological big data, "Catalyst role" in pharmaceuticals and complex diseases. Drug design time is reduced by 70%, and the success rate is increased by 10 times. According to data, the penetration rate of AI into the industry in 2018 was less than 5%, and it will reach 20% by 2026. Especially in the data-intensive biomedical industry, "embracing AI" is undoubtedly a hot exploration that is actively practiced from laboratories, clinicals, and production lines. Qiao Nan, chief scientist of Huawei Cloud Medical, mentioned in his speech that Huawei has released the Pangu large model in 2021 to provide services to industry customers. “Based on partners’ data, we work with customers to build basic large models to complete more industry scenario applications. ". Qiao Nan believes that China’s original drugs are relatively weak, and the intervention of AI is a good opportunity to systematically improve the level of my country’s pharmaceutical industry and achieve “overtaking in corners.” Practice has shown that AI pharmaceuticals can reduce drug design time by 70% and increase the success rate by 10 times. In the field of life and health, artificial intelligence has brought disruptive innovations to genomics, pharmaceuticals, and medical care. The pathogen identification and monitoring system of the Chinese Academy of Medical Sciences based on the genetic AI platform has increased the identification speed of pathogens by 5 times, saving 80% of labor costs; in Changhai Hospital, using generative AI technology, when the doctor completes the patient consultation, the medical record is also Generated in seconds; based on the Pangu drug model, the First Affiliated Hospital of Xi'an Jiaotong University discovered a super antibiotic in just a few weeks with very little manpower and capital investment. Qiao Nan said that large models will be combined with bioinformatics and organoid technology in the future to generate larger-scale high-throughput data. For example, Shenzhen Bay Laboratory is building a "BT+large model" high-throughput screening platform, which can reduce drug screening costs by 60% and achieve higher prediction accuracy. Mobile phone monitoring of abnormal behavior, AI breaks through Parkinson's disease diagnosis With the advent of the era of deep aging, neurodegenerative diseases such as Alzheimer's disease and Parkinson's disease have brought a heavy burden to society and families. Early identification, diagnosis and accurate Intervention is currently the best strategy to deal with these diseases. However, this is very difficult for specialist doctors. Wang Jian, director of the Department of Neurology at Huashan Hospital Affiliated to Fudan University, said that there are more than ten subdivisions of Parkinson's disease, and their intervention therapies are different, making accurate identification very difficult. However, AI will usher in a new era of diagnostic analysis of this disease. First of all, AI technology assists doctors in interpreting molecular images, which can significantly improve the diagnostic accuracy of Parkinson's disease. Molecular imaging of dopamine transporters and glucose metabolism networks in patients with Parkinson's disease shows characteristic changes, and the locations where related disease-causing proteins are abnormally accumulated in brain regions are also different. Wang Jian said that by cooperating with AI experts, large models can be trained to accurately identify characteristic changes and improve diagnosis accuracy. AI can also perform movement monitoring for Parkinson's disease patients through computer vision analysis of videos. For example, frozen gait is another characteristic of Parkinson's disease. AI analysis research on limb movement disorders based on video capture technology can accurately judge and quantitatively analyze patients' behavioral patterns, thereby better predicting the outcome of their disease. . "AI also makes 24-hour continuous monitoring at home possible. There are even studies on monitoring the respiratory signals of Parkinson's patients and extracting relevant features that can reflect the severity of the disease." Wang Jian said that not only in the field of AI research, At the industrial level, there are also professional companies applying AI technology to analyze the behavioral changes of Parkinson's disease. Related products have been approved for medical device license and are expected to be used in hospitals and home scenarios as soon as possible. Wang Jian revealed that the Parkinson's disease research team of the Department of Neurology of Huashan Hospital has launched an app called "Pawei" with more than 20,000 registered users. "This is the only Chinese app recommended on the website of the International Movement Disorder Society. ". Using this software, AI analysis can be performed on motor function videos and voice disorder audio uploaded by Parkinson's disease patients, and integrated into a key part of the Parkinson's disease chronic disease management platform, thereby enabling more data mining for real-world research. Multi-dimensional data exploration to solve the human body's "most complex disease" "There are currently no effective drugs to protect the heart on the market, probably because the heart is really too complex." The research team of Peter Ferdinandi, Vice President of Semmelweis University in Hungary I have been studying this field for more than 30 years, but the results have been limited. After introducing AI technology, they found a series of endogenous small RNA molecules that can protect the heart.
Why didn’t
Peter Ferdinandi, vice-chancellor of Semmelweis University in Hungary, discover it before? Ferdinandi explained that this is because the protection of the heart by these small RNAs is carried out in a molecular network, and its targets are quite complex. “Without the assistance of AI tools, it is difficult to discover from multi-omics data. ". Now, Ferdinandi's team has verified the cardiac protective effect of this small RNA in mouse and rat models. In the next two years, they will invest millions of euros to advance research into drug development. At present, they already have dedicated AI-driven drug development tools that can more efficiently find targets and design mRNA sequences for the development of oligonucleotide drugs for cardiovascular, tumor and other related indications. The human brain, known as the "three-pound universe", is the most complex organ in the world, and almost all brain diseases are unsolved mysteries. To this end, research on brain diseases has been included in brain science plans in many countries, and a massive amount of data has been generated so far. In 2018, the Zhangjiang International Brain Bank was established, which is the first Chinese multi-dimensional brain science data platform. Zhao Xingming, deputy director of the Institute of Brain-inspired Intelligence Science and Technology at Fudan University, said that the Zhangjiang Brain Bank currently has more than 520,000 cases of internationally shared data and 15,000 cases of domestic major brain disease data, and has identified 71 brain cell subtypes, covering 15 developmental stages and 63 brain areas. "Through AI data mining, we found that many brain disease drug targets are enriched in the small intestine, which will provide evidence for future new drug development and construction of comorbidity networks." Zhao Xingming revealed that they also found that neurological diseases and endocrine diseases, mental diseases and There is a strong comorbidity between gastrointestinal diseases.
