Premium
Power law analysis of the human microbiome
Author(s) -
Ma Zhanshan Sam
Publication year - 2015
Publication title -
molecular ecology
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 2.619
H-Index - 225
eISSN - 1365-294X
pISSN - 0962-1083
DOI - 10.1111/mec.13394
Subject(s) - population , biology , macroecology , spatial ecology , ecology , power law , spatial heterogeneity , population size , population variance , community , universality (dynamical systems) , stability (learning theory) , evolutionary biology , statistics , habitat , biodiversity , mathematics , demography , computer science , physics , quantum mechanics , machine learning , sociology
Taylor's (1961, Nature , 189:732) power law, a power function ( V = am b ) describing the scaling relationship between the mean and variance of population abundances of organisms, has been found to govern the population abundance distributions of single species in both space and time in macroecology. It is regarded as one of few generalities in ecology, and its parameter b has been widely applied to characterize spatial aggregation (i.e. heterogeneity) and temporal stability of single‐species populations. Here, we test its applicability to bacterial populations in the human microbiome using extensive data sets generated by the US ‐ NIH Human Microbiome Project ( HMP ). We further propose extending Taylor's power law from the population to the community level, and accordingly introduce four types of power‐law extensions ( PLE s): type I PLE for community spatial aggregation ( heterogeneity ), type II PLE for community temporal aggregation ( stability ), type III PLE for mixed‐species population spatial aggregation ( heterogeneity ) and type IV PLE for mixed‐species population temporal aggregation ( stability ). Our results show that fittings to the four PLE s with HMP data were statistically extremely significant and their parameters are ecologically sound, hence confirming the validity of the power law at both the population and community levels. These findings not only provide a powerful tool to characterize the aggregations of population and community in both time and space, offering important insights into community heterogeneity in space and/or stability in time, but also underscore the three general properties of power laws ( scale invariance , no average and universality ) and their specific manifestations in our four PLE s.