May 29, 2015 –– The human body is teeming with microbes that number in the billions. These microbes have caused quite a stir and been the subject of intense scrutiny in recent years as scientists continue to collect evidence of their role in a myriad aspects of human health. Advancing the field of microbiome research, though, has depended in part on more powerful and accurate statistical tools with which to analyze the data being collected – a hurdle cleared with a paper published today in the journal Microbial Ecology in Health and Disease.
In the paper, researchers from the Norwegian Institute of Public Health, the University of California (San Diego) and the National Institute of Environmental Health Sciences (NIEHS, NIH) in the US present a powerful new statistical methodology enabling more accurate, nuanced analysis of huge and highly complex microbiome data sets.
“This paper provides a general framework to analyze microbial data. We systematically describe various parameters related to microbiome data and develop a methodology in a principled way to compare two or more populations, such as the comparison of the abundance of different microbes in a baby delivered naturally and one delivered via C-section,” said the paper’s last author Shyamal Peddada, a senior investigator in the Biostatistics and Computational Biology Branch at the NIEHS (NIH).
The paper details a new statistical methodology termed ANCOM that allows researchers to compare microbial taxa abundance in two or more populations – including detecting trends over time in longitudinal or cross sectional studies, while adjusting for covariates if necessary. The authors describe ANCOM as a major advance over the existing methodologies at researchers’ disposal. In fact, in the paper, the authors draw attention to the fact that almost every statistical/computational methodology makes certain assumptions regarding the data being analyzed. Failure to satisfy these assumptions may lead to the use of an inappropriate methodology and thus potentially resulting in high false discovery rates and misinterpretation of the data. Accordingly, depending upon the question being answered, they caution against using some of the existing methods including t-test and ANOVA in microbiome studies.
“Given the proliferation of microbiome studies, there is an urgent need for the development of appropriate statistical methodology to properly analyze the data being collected,” Peddada said.
Tore Midtvedt, editor in chief of Microbial Ecology in Health and Disease, is excited by the possibilities he sees unfolding with the development of ANCOM. He points to its application in autism related microbiome studies as an example of how it will prove extremely valuable to researchers. “We have established that the gut bacteria of children with autism differs from that of typically developing children,” said Midtvedt.
“This new methodology will allow us to identify and describe temporal changes in the microbial composition in a child’s gut bacteria. But this is but one specific example. This paper represents a paradigm shift in how we evaluate complicated ecological systems.”
The paper, titled Analysis of Composition of Microbiomes (ANCOM): A novel method for studying microbial composition is freely available online in the journal Microbial Ecology in Health and Disease.
t: +1 561 935 3306
Microbial Ecology in Health & Disease is an Open Access peer reviewed Journal that publishes research on different human microbial eco-systems, to increase our understanding of their role in health and disease.
Tags: microbial composition, microbial ecology in health and disease, microbiome, National Institute of Environmental Health Sciences, National Institutes of Health, Norwegian Institute of Public Health, Shyamal Peddada, Tore Midtvedt