Scientists have proposed that the majority of microbes on Earth live in the vast, dark subsurface, but estimates of their numbers vary widely. Microbes in the subsurface might far outweigh the cells at the surface, or, they might just constitute a sizeable majority. During the past two decades, next-generation sequencing, which enables researchers to get a snapshot of a microbial community by sequencing all of the DNA in an environmental sample, has become more affordable and more widely used. Additionally, more scientists have undertaken explorations of the subsurface through coring and underground sampling. These advances are helping scientists to get a better handle on the numbers and diversity of cells in the subsurface.
In a new review in Nature Geoscience , DCO members Cara Magnabosco (Flatiron Institute, USA), Li-Hung Lin (National Taiwan University, Taiwan), Hailiang Dong (Miami University, USA), Malin Bomberg (VTT Technical Research Centre of Finland), Karsten Pedersen (Microbial Analytics Sweden), Thomas Kieft (New Mexico Institute of Mining and Technology, USA), and Tullis Onstott (Princeton University, USA), and colleagues, make new estimates of the number and biomass of cells living beneath the continents. The researchers compiled cell counts and next-generation sequencing data from sampling locations around the world and used advanced analysis techniques to scale local counts to a global approximation. They estimate that 200 to 600 octillion microbes (2 to 6x1029 cells) live in the continental subsurface.
Magnabosco embarked on this effort after completing her dissertation on microbes living deep within gold and diamond mines in South Africa. She wanted to know if the microorganisms she was seeing in her samples lived elsewhere in the subsurface. Onstott, her co-author and former advisor, had already begun compiling studies of subsurface diversity and microbial abundance for his book, Deep Life, so Magnabosco and her co-authors expanded the database to include data from more than 200 publications, with 3,800 measurements of cell concentrations in the continental subsurface.
The existing data sets had some significant gaps. Most data sets came from shallow locations in North America. But, by using data sets from all seven continents and taking into account surface temperature, global heat flow, depth, and rock type across the continents, the researchers were able to scale local biomass estimates to make a comprehensive estimate of biomass worldwide. To improve their accuracy, they performed cross validation, where they split their microbial data set into a “training set” and a “test set.” First they fed the training set into a software program to develop a model of global subsurface continental biomass (see the visualization below). Then they used the test set to assess and improve the model’s predictions.