Repository containing the scripts, the intermediate files, and the data used in the manuscript, "Island biogeography theory and the gut microbiome: why taller people tend to harbor more diverse gut microbiomes"
Before we can visualize and run regressions on our data, we need to generate the diversity metrics for each cohort, which is completed in the notebooks: HumanDiversity.ipynb and VertebrateDiversity.ipynb. These notebooks use the packages: pandas and qiime2. Each notebook:
- Converts our ASV or OTU table into a qiime feature table
- Summarizes the feature table
- Rarefies the feature table
- Quantifies the Simpson's Diversity for each sample in our cohort
- Converts the Simpson's Diversity returned by qiime from Simpson (1-D) to Simpson (1/D)
Now that we have Simpson's Diversity (1/D) calculated for our datasets, we can next visualize this data and run regressions. The American Gut cohort and the Arivale cohort are both analyzed in All_Human_Data.ipynb. A regression analysis in the Arivale dataset to examine the effect of health on regressions is in sick_arivale_participants.ipynb. The vertebrate datasets: Godon et al., Song et al., and Groussin et al. are analyzed in All_Vertebrate_Data.ipynb. Each notebook:
- log transforms height or mass and Simpson's Diversity
- For our human datasets:
- we add in relevant covariates to our datasets for inclusion in regressions
- Standardize the data
- Plot using the Seaborn package
- Run OLS regressions with the Statsmodels package