Traits track taxonomy

A novel technique based on isotope analysis shows that, compared to ecosystem type, evolutionary history explains more variation in bacterial growth traits along an elevation gradient. This knowledge could help move microbial ecologists toward improved predictive models of soil processes.

Bacteria and Archaea account for the vast majority of genetic diversity of life on Earth1. Studying such an enormous trove of microbes — and the specific roles they play — has been a major challenge for microbial ecologists. In this issue of Nature Ecology & Evolution, Morrissey et al.2 tackle diverse microbial communities with a powerful new isotope technique in soils of the southwestern United States. The approach reveals that evolutionary history explains much of the variation in microbial growth traits.

The link between specific organisms and the functioning of ecological communities has fascinated ecologists for centuries. So-called ‘structure–function’ relationships are the linchpin needed to predict ecosystem functioning in a changing climate, for example3, and here the authors aimed to test whether key determinants of microbial functioning were similar across different environments. One hypothesis is that phenotypic characteristics, or traits, are more similar for more closely related microbes4. Such consistency would be convenient because ecologists could predict ecosystem functioning on the basis of which microbial taxa are present. If traits reflect evolutionary history, taxonomic information might be sufficient to build a predictive model of ecosystem processes, such as carbon cycling. And advances in high-throughput DNA sequencing mean that taxonomic information is easy to get for almost any microbial community.

What’s harder to get is accurate data on microbial traits and how traits vary under different environmental conditions, especially for intact communities in the field. Previous studies have measured traits on pure cultures of microbes5, but this work is painstaking and subject to methodological limitations. There are some emerging approaches for analysing the traits of whole microbial communities, but these measurements are low in taxonomic resolution6. Using a new technique known as quantitative stable isotope probing, or qSIP, Morrissey et al. measured growth and resource assimilation traits of intact bacterial communities in soils across an elevation gradient in Arizona, United States. The method requires an ultracentrifuge and some clever math combined with modern DNA sequencing tools, but what results is a detailed profile of growth and carbon assimilation rates for every resident bacterial group that can be sequenced from a given soil.

With sites along the elevation gradient ranging from dry grassland to coniferous forest, ecosystem type was a key variable that represented substantial environmental variation in the study design. Evolutionary history — represented as taxonomic assignment to groups like phylum and family — was the other main experimental variable.

A nested analysis of variance gave a clear result. Taxonomy, and the evolutionary history it represents, explained a much larger fraction of the variation in bacterial traits than ecosystem type (Fig. 1). Whereas ecosystem explained 20% of the variation at most, taxonomy explained up to 65%. Family and bacterial phylotype (roughly the species level) accounted for most of the variation attributed to taxonomy.

Fig. 1: Using a new technique with isotope labels, Morrissey et al.2 show that evolutionary history explains more variation in traits of soil bacteria than soil conditions in ecosystems across an elevation gradient.

The link between taxonomy and traits like growth rate and carbon assimilation could be useful in modelling soil processes under a changing climate. Coloured areas approximate the proportion of variance explained by ecosystem type (green) or taxonomy (blue) for bacterial growth rates (a) or carbon assimilation rates (b) with added growth substrates. Unexplained variation is shown in grey. Isotopically labelled water and glucose are taken up and shown inside soil bacteria.

The taxonomic level accounting for the most variance can vary widely by trait7. Complex traits like methane production that require many genes are often deeply conserved, for example at the order level. More simple traits — like the assimilation of carbon substrates — involve fewer genes and are generally conserved somewhere from the genus level to phylotype level in culture-based studies. The analysis by Morrissey et al. suggests that rates of growth and glucose assimilation are also relatively simple traits that are not deeply conserved for bacteria growing in soil.

Growth and carbon assimilation traits are important in microbial communities. Growth rate reflects the physiological balance between building biomass and expending resources to survive. Carbon assimilation rate gives information about potential resource use and competitive ability. The holy grail for structure–function researchers is to apply trait data like Morrissey et al. collected to predict ecosystem functioning. Still, there is a long way to go. The empirical data would have to be incorporated into a mathematical model, and such a model would require much more information.

Although they analyse soils from distinct ecosystem types, Morrissey et al. did not address how growth traits respond to abiotic conditions like temperature and soil moisture. Data of that kind are needed to build dynamic models in which traits and functions change realistically with climate and other drivers. Another unresolved issue for prediction is how to represent the taxonomic diversity of microbes in an ecosystem model. The authors nicely demonstrate that different bacterial families and phylotypes have distinct traits that relate to carbon and nutrient cycling. But even the most sophisticated ecosystem models only include a handful of different microbial groups8, not the hundreds found in the Arizona soils. Modellers need a scheme to lump this taxonomic diversity into a manageable number of groups or to represent it through continuous distributions. Overall, progress on prediction will require a lot more cross-talk between microbial empiricists and ecosystem modellers8.

Even though accurate predictive models may elude us, the results from Morrissey et al. still stand out as an important contribution. Ecologists have long known that traits correspond to taxonomy in macroscopic organisms like trees and birds9,10. Pine trees have needles regardless of whether they grow in Canada or Central America. Yet finding the same pattern with microbes was not guaranteed. Bacteria can evolve and exchange genes rapidly11,12, so bacterial traits might have had very short evolutionary histories. Any differences among taxa could have been washed out quickly, leaving the soil environment as the dominant force determining bacterial traits.

But that’s not the case. Thanks to Morrissey et al., we now know that a Bacillus keeps its high growth rate just as a pine tree keeps its needles, no matter where it grows.


  1. 1.

    Pace, N. R. Science 276, 734–740 (1997).

    CAS  Article  Google Scholar 

  2. 2.

    Morrissey, E.M. et al. Nat. Ecol. Evol. (2019).

  3. 3.

    Bardgett, R. D., Freeman, C. & Ostle, N. J. ISME J. 2, 805–814 (2008).

    CAS  Article  Google Scholar 

  4. 4.

    Philippot, L. et al. Nat. Rev. Microbiol. 8, 523–529 (2010).

    CAS  Article  Google Scholar 

  5. 5.

    Dolan, K. L., Peña, J., Allison, S. D. & Martiny, J. B. H. PLoS One 12, e0174472 (2017).

    Article  Google Scholar 

  6. 6.

    Fierer, N., Barberan, A. & Laughlin, D. C. Front. Microbiol. 5, 614 (2014).

    Article  Google Scholar 

  7. 7.

    Martiny, J. B. H., Jones, S. E., Lennon, J. T. & Martiny, A. C. Science 350, 649 (2015).

    CAS  Article  Google Scholar 

  8. 8.

    Wieder, W. R. et al. Glob. Biogeochem. Cycles 29, 1782–1800 (2015).

    CAS  Article  Google Scholar 

  9. 9.

    Neyret, M. et al. Ecol. Evol. 6, 5674–5689 (2016).

    Article  Google Scholar 

  10. 10.

    McGill, B. J. Am. Nat. 172, 88–101 (2008).

    Article  Google Scholar 

  11. 11.

    Lenski, R. E. ISME J. 11, 2181–2194 (2017).

    CAS  Article  Google Scholar 

  12. 12.

    Gogarten, J. P. & Townsend, J. P. Nat. Rev. Microbiol. 3, 679–687 (2005).

    CAS  Article  Google Scholar 

Download references

Author information



Corresponding author

Correspondence to Steven D. Allison.

Ethics declarations

Competing interests

The author declares no competing interests.

Rights and permissions

Reprints and Permissions

About this article

Verify currency and authenticity via CrossMark

Cite this article

Allison, S.D. Traits track taxonomy. Nat Ecol Evol 3, 1001–1002 (2019).

Download citation


Quick links

Nature Briefing

Sign up for the Nature Briefing newsletter — what matters in science, free to your inbox daily.

Get the most important science stories of the day, free in your inbox. Sign up for Nature Briefing