Research notebook for manuscript on taxonomic bias and differential-abundance methods
Compare impact of bias in HMP stool and vagina samples
This report compares gut and vaginal microbiome profiles from the Human Microbiome Project to consider how the different ecological dynamics in each sample type might modulate the potential for bias to distort FCs in species proportions.
Explore bias sensitivity of Vieira-Silva et al (2019) QMP analysis
Analysis of Leopold and Busby (2020) for manuscript case study
This document performs analysis of the data from @leopold2020host to support the fungal case study in the manuscript.
Analysis of Lloyd et al (2020) for manuscript case study
Perform calculations for the Lloyd et al (2020) case study.
Perform Leopold and Busby (2020) regression analysis with and without bias correction
We perform the regression analysis of Leopold and Busby (2020) with and without bias correction, finding that there is negligible impact of bias correction on the results.
Perform Leopold and Busby (2020) priority-effects analysis with and without bias correction
MOMS-PI: Regression on diversity
Analyze the impact of bias on a DA analysis of relative abundances versus diversity.
MOMS-PI: Key dynamics
Illustrate the basic problem
Illustrate calibration types in the mock communities from Brooks et al (2015)
Evaluate robustness example
Evaluate the robustness of an estimate and inference to bias.
MOMS-PI patient exploration and analysis of fold changes
Simulate regression example
Simulate an example of how bias impacts a regression analysis.
Hypothetical example of spurious fold changes in proportions in the vaginal microbiome
This doc develops a hypothetical example in which bias leads to a spurious differential abundance result. It is based on a real case of bias measured among three species of vaginally-associated bacteria.
Illustration of a spurious differential abundance result due to bias
Create illustration of a spurious differential abundance using the Leopold and Busby (2020) dataset.
