In the frontier exploration of biological science, protein, as the basic building block of life, is of great importance. Recently, a breakthrough achievement of Shanghai Jiao Tong University has attracted widespread attention. Distinguished Professor Hong Liang and his team have successfully launched a large protein design model called Venus, an innovative tool that can accurately predict the change of protein from sequence to function, realize the "directed evolution" of protein functional design, and transform the traditional "needle in a haystack" method of protein research into an efficient and accurate custom design.

The application of the Venus model is not only at the theoretical level, but also the industrialization process of many proteins designed by it. For example, a PET-degrading enzyme with excellent activity, strong stability, and pure product, as well as a highly active alkaline phosphatase for ultrasensitive detection and diagnosis, are examples of the successful application of the Venus model. The launch of these protein products heralds a revolutionary change in the field of protein engineering.

This innovative achievement of the Hongliang team benefits from their deep integration of artificial intelligence technology with protein design and modification. The team built the world's largest protein dataset, which contains billions of protein sequences, covering a wide range from conventional surface organisms to extreme environment microorganisms, with hundreds of millions of sequences also functionally tagged. The Venus model, trained on this large dataset, can efficiently and accurately predict and design the function of proteins, transforming the protein production process from "slow trial and error" to "precise design".

In practice, the Venus model has shown amazing capabilities. Hongliang's team cooperated with Kinsey Pharmaceutical, a leading company in the field of growth hormone in China, to use the Venus model to modify the alkali resistance of single-domain antibodies. In less than a year, the alkali resistance of ordinary single-domain antibodies was increased by 5000 times, saving tens of millions of yuan in costs for Kinsey Pharmaceutical. This achievement not only achieved the scale-up production of multiple batches of 0 liters, but also became the world's first protein product designed and produced on a large scale, marking the great potential of the Venus model in industrial application.

The Venus series of models has also made important breakthroughs in the field of in vitro diagnostics. The team successfully optimized an alkaline phosphatase (ALP) to make its molecular activity 200 times higher than that of the products of leading international companies. This achievement is of great significance for the diagnosis of hypersensitivity detection, and the early detection of diseases such as myocardial infarction and Alzheimer's disease will become more accurate and efficient. At present, the transformed ALP has entered the stage of scale-up production with a scale of 0 liters, marking the success of the Venus series model in industrial transformation.

The Hong Liang team also released VenusMutHub, which is the first small-sample dataset of protein mutations for real-world application scenarios. The team also proposed a zero-sample prediction evaluation standard for protein function, and the related research has been published in the Journal of Pharmaceutical Sciences. These results not only enrich the data resources in the field of protein design, but also provide more scientific and accurate evaluation standards for future research.

With the launch of the Venus model and the VenusMutHub dataset, the Hongliang team has made remarkable achievements in the field of protein design and engineering. These innovations not only promote the development of biological sciences, but also bring new hope and opportunities to the fields of medical health and environmental protection.