The liquid under the lab bench looks like leaf litter soup. Vials of brown, ochre, and reddish dust, extracted from pollen, clutter the surface above. Beside them, a foil-covered beaker contains a cloudy goo derived from horny goat weed.

To biologist Jing-Ke Weng, that mess is a rich source of new medicines: drugs to treat cancer, perhaps, or blood ailments.

Many drug discovery experts would warn him away from such dreams. They’d point out that the biggest pharma companies in the world have spent decades, and millions, trying to make drugs from plants. They’d tell him it’s just too hard.

Weng, though, isn’t worried. At 34, he’s already helped map the molecular factories that produce some of the most important plant chemicals around. And now that he’s working at the Whitehead Institute for Biomedical Research, he’s determined to harness peanut skins and Kava Kava root — and the other strange vegetation in his seventh-floor greenhouse — to treat human disease.

The idea is not new. Some of our best-selling commercial drugs come from plants: aspirin from willow bark, the cancer drug Taxol from the Pacific yew tree, the leukemia medication vincristine from a pinkish jungle flower called the rosy periwinkle. And last year, Chinese chemist Tu Youyou shared the Nobel Prize in Physiology or Medicine for her research on artemisinin, a malaria drug extracted from a bush called sweet wormwood.

But this type of botanical sleuthing is an anomaly in the biotech hotbed of Kendall Square.

Most large pharmaceutical companies phased out their screening of natural products in the 1980s. It was expensive, and eureka moments often turned sour: The researchers kept rediscovering molecules that had already been commercialized.

Academic labs and small startups have taken over some of that research, but they’ve focused mostly on bacteria and fungi, because in those microorganisms, the genes that code for medicinal compounds tend to stick together.

Not so with plants, said Gregory Verdine, a chemist at Harvard and founder of the natural products company Warp Drive Bio. “If you go into plants, you add another layer of complexity to what’s already incredibly complex,’’ he said.

Weng, though, finds the complex fascinating. Most scientists are loath to use the word “magical’’— but that is Weng’s go-to adjective when describing his plants.

The conviction began when he was a kid in Hangzhou, China. If he got sick, his mother would walk out of their first-floor apartment to collect a potful of what he translates as “fishy-tasting grass.’’ In those days, it grew everywhere, like a weed. His mother would bring it to a boil on the stove for 20 minutes, and then scoop out the soggy blades, serving him the blackish, smelly liquid that remained.

“It doesn’t look good and it doesn’t taste good, but I like it, because it really helps when you feel bad,’’ he said.

As an undergraduate, Weng focused on neuroscience research. But he didn’t like having to kill mice for work. And so, for his PhD at Purdue, he returned to his love of plants and began to work on the chemistry of wood.

“He was the most creative student I have ever had and probably will ever have,’’ said Clint Chapple, his PhD supervisor. “Most PhD students graduate with one, two, three scientific papers that they’ve led. Jing-Ke published 13 papers out of his PhD.’’

Weng may have been a superstar of a student, but the job offer from the Whitehead Institute surprised him. It’s a biomedical center full of cancer biologists; he studied how plants made rigid cell walls.

But he was unfazed. He knew how to break apart assembly lines of enzymes — and those cellular factories produced an almost infinite catalogue of strange chemicals, many of them medicinal.

He began to seek them out.

From a lecturer at a natural history museum, he learned that Micronesian tribes used the calming Kava Kava root for rituals. Soon after, he was unpacking live plants from a nursery in Hawaii. At a high school science fair, Weng’s father saw an experiment suggesting that peanut skins could help patients with low platelet counts. He mentioned it to his son; Weng bought 11 pounds of peanut skins from China that are now soaking in ethanol under a lab bench.

One of his post-docs — a surfer — pulled a wrack of red algae from a tide pool in California, having heard it harbored cancer-fighting compounds. Now they’re preparing to sequence it.

When he gets a plant, the first thing Weng does is to destroy it. He separates leaf from stem from root. Then each part is mashed up. The genetic material is extracted and sequenced, while another machine identifies the plant’s chemicals, so that he can link up those “magical’’ molecules to the genes that code for them — and hopefully mass-produce them in genetically engineered microbes.

Weng at times uses Amazon.com to initially gauge the potency of a plant, looking at the comments from people who have ordered herbal supplements. “People actually are [reporting] the type of response they’ve observed from their own body. It’s free information,’’ he explained.

It’s hard to say whether any of the shoots in his greenhouse — the tiny reddish green plant, or the flower that looks like the frilly hem of a flamenco dress — will yield an FDA-approved drug.

But as Weng walks among his beakers clouded with bacteria and his vials of pollen, he doesn’t look nervous.

Eric Boodman can be reached at eric.boodman@statnews.com. Follow him on Twitter @ericboodman. Follow Stat on Twitter: @statnews.