From 62eabcfb778982001920620ff84be918b674d6c9 Mon Sep 17 00:00:00 2001 From: Michael McLaren Date: Tue, 16 Nov 2021 07:15:40 -0500 Subject: [PATCH] Edit abundance-measurement section --- abundance-measurement.Rmd | 10 +++++----- 1 file changed, 5 insertions(+), 5 deletions(-) diff --git a/abundance-measurement.Rmd b/abundance-measurement.Rmd index 4d00b7b..80047d0 100644 --- a/abundance-measurement.Rmd +++ b/abundance-measurement.Rmd @@ -38,11 +38,11 @@ is the average or mean efficiency of cells in the sample. ## Relative abundances (proportions and ratios) -We distinguish between two types of species-level *relative abundances*---abundances in which the absolute scale is not intrinsically meaningful. +We distinguish between two types of species-level *relative abundances*. The first type consists of the *ratios* among the abundances of two or more species, such as a ratio of 5:4:2 among three species. -The second type consists of the *proportion* of an individual species, equal to its abundance divided by that of the chosen root taxon (e.g. Prokaryotes). -Species' ratios and proportions constitute distinct views of relative abundance and hence what it means for relative abundances to change across samples. -Although proportions are dominant in microbiome DA analysis, researchers increasingly adopt the ratio view via analyses adapted from the statistical framework of Compositional Data Analysis (CoDA). +The second type consists of the *proportion* of an individual species, equal to its abundance divided by that of a particular root taxon (e.g. Prokaryotes). +DA analyses of proportions consider how the proportion of a species changes indepently of other species, whereas DA analyses of ratios consider the conserted change in relative abundance among two or more species. +Although proportion-based analysis is dominant, microbiome researchers increasingly use ratio-based analyses adapted from the statistical framework of Compositional Data Analysis (CoDA). The different manner in which taxonomic bias affects ratios vs. proportions (@mclaren2019cons) forms the basis in subsequent sections for understanding the extent to which its effects cancel in DA analysis. The standard measurement of the proportion of species $i$ in sample $a$ is given by the ratio of its read count to the total, @@ -96,7 +96,7 @@ rm(svg_path, fig_path) **Difference between species proportions and species ratios:** Consistent taxonomic bias at the species level creates a consistent fold error in the ratios among species, but a varying fold error in the proportions of individual species. But a proportion is, after all, just a ratio between two taxa in which the denominator taxon is the entire taxonomic domain under study. -The difference in behavior from that of ratios between species arises because the efficiency of higher-level taxa, which consist of the aggregate of many species, varies across samples with the changing relative abudances among its constituent species. +The difference in behavior from that of ratios between species arises because the efficiency of higher-level taxa, which consist of the aggregate of many species, varies across samples with the changing relative abundances among its constituent species. ## Absolute abundances (cell densities)