As a high school senior, Bing Zhu was not the most dedicated student. Instead of studying for college entrance exams, he spent his time playing and learning strategies for Go, an ancient Chinese board game. It proved a smart career investment: Zhu credits the strategic thinking he learned with an ability to plot his way through the most tangled biological puzzles.

In Go, Zhu says, you must envision your strategy and always think ahead of your opponent. “It’s like when we do an experiment—sometimes we get a surprise and you have to reevaluate your strategy.” His ability to think ahead of the curve led him to invalidate one model of how epigenetic information is inherited.

In college at Zhejiang University, Zhu had less time for Go as he gravitated toward classes in biochemistry, molecular biology, and genetic engineering. He went on to study genetic engineering, getting his master’s at China National Rice Research Institute and his Ph.D. at the Shanghai Institute of Plant Physiology, where his research centered on inserting and removing genes from plants and bacteria.

But it was during his first postdoc, studying under Jean-Pierre Jost at the Friedrich Miescher Institute in Basel, Switzerland, that Zhu discovered an interest in the burgeoning field of epigenetics—heritable changes to gene expression that are reflected not in the DNA sequence itself but in modifications to the DNA or histone proteins that package the DNA. Such modifications help determine whether a gene is expressed. Moreover, the pattern of modification is passed from generation to generation during cell divisions, making it truly epigenetic.

“I was trying to understand how genes get regulated,” Zhu says. “Why are some genes expressed in some cells but not others?” The expression of different genes by different cells is what allows complex, multicellular organisms to exist, he explains, by doling out different jobs to different cells and organs.

Zhu moved to the University of Medicine and Dentistry of New Jersey for his second postdoc, where he worked with Danny Reinberg, an HHMI investigator who is now at New York University. There, Zhu delved more deeply into gene regulation, identifying new protein complexes involved in histone modifications. But more important than the research, Zhu says, he learned how tactical planning could help him thrive at the lab bench. “Danny gave his postdocs a lot of guidance and scientific freedom simultaneously, allowing us to fail and helping us to succeed,” Zhu says. “And I learned that success in science all comes back to the game of Go. It’s strategic thinking that’s most important.”

In 2006, Zhu returned to China to start his own lab at the National Institute of Biological Sciences, Beijing. He’s been studying the inheritance of epigenetic information ever since, once again focusing on histone modifications.

He turned his attention to one highly attractive model of epigenetic inheritance, which proposed that histone modifications are faithfully copied within each nucleosome. The model suggested that each nucleosome—made up of eight histone proteins—split into two groups of four to be copied. The two resulting histones would each have four of the original proteins and four new proteins. But Zhu noticed that the model relied on two assumptions. First, that the histone modifications in each nucleosome must be symmetric, and second, that the nucleosome cores separated during the process of DNA replication.

By systematically testing these assumptions and finding them untrue, he has proven the overall theory incorrect. He’s also gone on to show that exact copying isn’t always necessary for histone modifications to pass from mother to daughter cell. “Histone modifications are more or less like a buffer and can tolerate certain fluctuation without changing the eventual outcome,” he says. “Nature is smarter than we used to think—our bodies know that some fluctuation can happen and you don’t have to respond to it in a super precise way.”

Now, he’s working to come up with a theory on how histone modifications are inherited. “Because we basically screwed up a model, we need to establish something else,” Zhu says. More than the answers, however, it’s the process of obtaining them that he loves most. “For me, the most entertaining thing is the experimental design. In science, like in Go, winning is the by-product of good design.”