Genomics pioneer Jun Wang on his new AI venture

David Cyranoski in Nature:

Jun Wang is one of China’s most famous scientists. Since joining the genome-sequencing powerhouse BGI when it started up 16 years ago, he has participated in some of its biggest accomplishments. These include sequencing the first genome of an Asian person¹, the giant panda² and the human gut microbiome³, as well as contributions to the Human Genome Project. Wang has led BGI since 2007 (when it stopped using the name Beijing Genomics Institute and moved its headquarters to Shenzhen). But on 17 July, the institute announced that he will give up that position to pursue research into artificial intelligence (AI).

What is the concept behind your AI project?

Basically, I am just trying to feed an AI system with masses of data. Then that system could learn to understand human health and human life better than we do. The AI will try to draw a formula for life. Life is digital, like a computer program — if you want to understand the results of the programming, how the genes lead to phenotypes, it is sufficiently complicated for you to need an AI system to figure out the rules. The AI system will basically consist of two components. The first is the big supercomputing platforms. We already have access to those through cloud computing and supercomputing centres. These will run or devise algorithms that look for relationships between genes, lifestyle and environmental factors, and predict phenotypes. The other thing is big data. We want to have data from one million individuals. And we want the data to be alive, in the sense that they can update their phenotype information at any time point. Other big computing companies, such as Google, could eventually do this, but we want to do it first. And we have the experience with the big data.

More here.