Scientists call for global cooperation in creating AI virtual cells

Science Daily (Reporter Zhang Jiaxin) The scientific team of Stanford University, Genentech Pharmaceuticals, and the Chan-Zuckerberg Foundation pointed out in a paper published in the latest issue of Cell that, given the latest advancements in artificial intelligence (AI) and the availability of large-scale experimental data on human biology, the scientific community has been presented with an unprecedented opportunity to use AI to create virtual human cells. These cells will be able to represent and simulate the precise behavior of human biological molecules, cells, and ultimately tissues and organs.

This synthetic cell model will enable people to have a deeper understanding of the complex interactions of chemistry, electricity, mechanics, and other forces and processes involved in the functioning of healthy human cells, and reveal the underlying causes of diseases that lead to cellular dysfunction or death.

The paper suggests that perhaps even more exciting is that AI virtual cells will allow scientists to conduct experiments on computers without the need for live cells and organisms. This ability will expand human understanding of human biology, accelerate the development of new drugs, and the emergence of new therapies.

With the help of AI virtual cells, cancer biologists can simulate how certain mutations transform healthy cells into malignant cells, microbiologists can predict the impact of viruses on infected cells and even the host body, and doctors can test treatment methods on the patient’s “digital twin”, thus accelerating the long-awaited goal of faster, more economical, and safer personalized medical care.

However, the paper states that in order to succeed, AI virtual cells need to achieve three goals: first, they must enable researchers to create universal representations across species and cell types; second, they must accurately predict cell function, behavior, and dynamics, and understand cellular mechanisms; finally, AI virtual cells must be able to conduct experiments on computers to test hypotheses and guide data collection, thus expanding the ability of virtual cells at a much lower cost and speed than currently.

The paper argues that AI has ushered in an era in the scientific community with predictable, generative, and searchable tools. However, achieving AI virtual cells is by no means an easy task and requires global open scientific cooperation in fields such as genetics, proteomics, and medical imaging.

The paper proposes close collaboration between global stakeholders, including academia, industry, and non-profit organizations, to jointly create the world’s first AI virtual cell. However, the paper also warns that any work on AI virtual cells is premised on the model being unrestrictedly open to the entire scientific community.

Share to let more people see

Source: People’s Daily Online

Author: Science and Technology Daily