Shanghai Jiao Tong University and Tianwu Tech Unveil Venus-Pod: The World's Largest Protein Dataset with 9 Billion Sequences

On March 22, byShanghai Jiao Tong UniversityUnitedMatwings TechnologyHosted the "Protein Function Prediction Venus Series Model Release and Industry Cooperation Summit".

At the meeting, Shanghai Jiao Tong UniversityLiang HongThe professor's team released the latest results, combining AI with protein design and modification, establishingThe World's Largest Protein Dataset；

Models trained on this dataset can accurately and efficiently predict and design protein functions, transforming protein production from "slow trial and error" to "highly efficient precision design."

It is reported that, with the support of industry-leading automated equipment, this achievement has been industrialized and implemented. For example,Highly active, strongly stable, and highly pure PET-degrading enzyme, and highly active alkaline phosphatase for ultrasensitive detection and diagnosisetc.

The World's Largest Protein Sequence Dataset

The protein sequence dataset Venus-Pod (Venus-Protein Outsize Dataset) released this time contains nearly 9 billion protein sequences and includes hundreds of millions of functional tags.

This model is the world's largest dataset in terms of data scale and contains the most functional annotation tags, as well as the 2.1 billion protein sequences used to train another industry-renowned model—the US ESM-C model.4x Volume。