This Github page deposits quality trimming, MOTHUR preprocessing, and downstream analysis scripts for manuscript "Nemolizumab treatment normalizes dysregulated microbiome in lesional prurigo nodularis"
- quality_trimming.sh: conduct quality trimming on the raw fastq sequences using fastqc to trim bases with quality scores less than 25
- mothur.sh: mothur pipeline to process paired trimmed fastq.gz data files, output otu table as well as taxonomy assignments
- data_preprocessing.R: process mothur outputs, conduct low abundance taxon removal
- descriptive.R: conduct descriptive analysis on genera and phylum in our data, generate figures for Figure 2
- shannon.R: calculate shannon diversity for each sample using rarefication, test for its significance with clinical variables
- bray-curtis.R: calculate bray-curtis distance between samples using rarefication, conduct dimension reduction using NMDS, and test for significance between different skin types and batches
- DE_{taxon}_{method}.R: conduct differential abundance/expression analysis
taxon: taxonomy-level of phylotype unit, genus-/phylum-level are used in this study
method: method used for differential analysis, ANCOMBC, DESeq2, and limma-voom are used - plot_DE_{genus, phylum}.R: scripts to plot ANCOMBC results and generate figures for Figure 4
- mlr_{baseline, nrsw4}.R: machine learning scripts to classify baseline lesional/non-lesion skin types and predict nemolizumab response status using microbiome and demographic data.