Integration of microbial ecology and statistics: a test to compare gene libraries

Abstract

Libraries of 16S rRNA genes provide insight into the membership of microbial communities. Statistical methods help to determine whether differences in library composition are artifacts of sampling or are due to underlying differences in the communities from which they are derived. To contribute to a growing statistical framework for comparing 16S rRNA libraries, we present a computer program, integral -LIBSHUFF, which calculates the integral form of the Cramér-von Mises statistic. This implementation builds upon the LIBSHUFF program, which uses an approximation of the statistic and makes a number of modifications that improve precision and accuracy. Once integral -LIBSHUFF calculates the P values, when pairwise comparisons are tested at the 0.05 level, the probability of falsely identifying a significant P value is 0.098 for a study with two libraries, 0.265 for three libraries, and 0.460 for four libraries. The potential negative effects of making the multiple pairwise comparisons necessitate correcting for the increased likelihood that differences between treatments are due to chance and do not reflect biological differences. Using integral -LIBSHUFF, we found that previously published 16S rRNA gene libraries constructed from Scottish and Wisconsin soils contained different bacterial lineages. We also analyzed the published libraries constructed for the zebrafish gut microflora and found statistically significant changes in the community during development of the host. These analyses illustrate the power of integral -LIBSHUFF to detect differences between communities, providing the basis for ecological inference about the association of soil productivity or host gene expression and microbial community composition.

FIG. 1.

A partial plot of C_X and C_XY that shows the differences found using the integral form and approximation form of Cramér-von Mises statistic as a function of the size of the distance window used to identify singleton sequences. Plot was generated by using comparison between soil 16S rRNA gene libraries of McCaig et al. (20). Vertical bars denote the locations where the approximation form calculates C_X and C_XY as implemented in LIBSHUFF with a ΔD value of 0.01. The integral form implemented in ∫-LIBSHUFF calculates C_X and C_XY continuously over the entire range of observed distance values.

FIG. 2.

Comparison of the taxonomic diversity in Wisconsin from 1997 (n = 139) and 2000 (n = 128) and improved (n = 138) and unimproved (n = 137) Scottish soil clone libraries as reported in original publications for most abundant phyla (A) and using collector's curves for a distance of 0.20 (B). OTUs, operational taxonomic units.

FIG. 3.

Collector's curves constructed using distance-based OTU and richness for distances of 0.03 and 0.20, assuming that GenBank accession numbers represent the order in which the sequences were collected. The collector's curves are divided into six sections, representing the six libraries that Rawls et al. constructed: 6 dpf, conventionally raised (A; n = 362), 6 dpf, conventionally raised (B, n = 239), 10 dpf, conventionally raised (C; n = 35), 20 dpf, conventionally raised (D; n = 103), 30 dpf, conventionally raised (E; n = 76), and adult, conventionally raised (F; n = 167). OTUs, operational taxonomic units.