This article was originally posted on RealClearScience.

Many bacteria are notoriously picky eaters. The microbe that causes leprosy, for instance, cannot be grown in a test tube. Instead, researchers must culture the microbe in armadillos or on the footpads of mice. Other bacteria without foot fetishes can still be difficult to culture, requiring a long and complex recipe of various nutrients, ions, and vitamins. As a result, it is simply impossible for microbiologists to grow some 99% of bacteria in the laboratory.

Ironically, technological advances in molecular biology have greatly outpaced those of basic microbiology. Thanks to molecular biology, bacteria can be easily identified with just a tiny amount of DNA. A sample of ocean water or soil may contain hundreds of different microbial inhabitants. By extracting all of the DNA and sequencing a particular gene (present in all bacteria) known as 16S rDNA, scientists can identify many of the bacteria present. However, microbiologists are left frustrated by the fact that, though the bacteria are identifiable, they remain unculturable in the lab and hence recalcitrant to further study.

But now, thanks to an international team of researchers led by Matthew Oberhardt and Eytan Ruppin, that may be about to change. If successful, their research could represent nothing less than a revolution in basic microbiology.

The team created a gigantic database of more than 20,000 pairings of microbes and culture conditions. Then, they created an algorithm, called GROWREC, that predicts the growth requirements of a bacterium using only its 16S rDNA sequence. Finally, the team tested how 61 different bacteria grew in their predicted culture conditions:

The authors found that the predicted culture conditions for 26 (red) of the bacteria were confirmed by existing literature; 29 (blue) more were confirmed when the authors grew the bacteria in their own labratories. Only six (green) of the 61 bacteria failed to grow.

While the authors did not attempt to grow “ungrowable” bacteria using their algorithm, they believe that they have shown a proof-of-principle. Additionally, they created an online tool for any microbiologist to use. With time, the researchers should be able to collect enough data to determine how accurate and useful their algorithm actually is.

If their tool actually is as powerful as they believe it to be, the team’s research could have an enormous impact not only on basic microbiology but biotechnology as well. Some of the literally millions of as-yet-uncultured microbes on our planet likely produce novel antibiotics or possess metabolic pathways useful for everything from creating biofuels and bioplastics to degrading toxic compounds. Harnessing the untapped potential of these microorganisms often begins by growing them in the lab.

Source: Matthew A. Oberhardt, Raphy Zarecki, Sabine Gronow, Elke Lang, Hans-Peter Klenk, Uri Gophna & Eytan Ruppin. “Harnessing the landscape of microbial culture media to predict new organism–media pairings.” Nature Communications 6, Article number: 8493. Published: 13-October-2015. doi: 10.1038/ncomms9493