The total number of microbial species on the planet Earth is two orders of magnitude smaller than the consensus estimate, according to a data analyses performed by John Bunge, chair of social statistics and associate professor of statistical science at Cornell University.
Bunge and his research colleagues collaborated with database consultants at the Cornell Center for Advanced Computing (CAC) and used their high-performance database servers to create a more accurate estimate of the biodiversity of microbial populations.
Bunge used a combination of statistical techniques – parametric and nonparametric, frequentist and Bayesian, and abundance- and incidence-based – to generate this new insight. Typically diversity is defined as the number of taxonomic groups such as species or genera, but the same statistical techniques can be used to estimate the number of individuals within a population.
Bunge's research team investigates microbial diversity, both in specific environments – such as sub-oceanic hydrothermal vents, Arctic ice melt, and beach sands – and over the entire biosphere. 'CAC deals with this scenario very well,' said Linda Woodard, a research specialist focused on the simulation modelling of ecological systems and database design. 'Each run is independent and can be done in parallel. While researchers can only run one data set at a time on a desktop computer, we can process multiple data sets simultaneously with our high-performance database servers.'
The Cornell Center for Advanced Computing receives support from the Cornell University, the National Science Foundation, DOD, USDA, and members of its corporate programme, including Pfizer, Boeing, Corning, and Northrop Grumman.
