Analysis scripts for "Novel multiparameter correlates of Coxiella burnetii infection and vaccination identified by longitudinal deep immune profiling."
https://doi.org/10.1038/s41598-020-69327-x https://www.nature.com/articles/s41598-020-69327-x
Includes scripts for SpadevizR/cluster matching analysis of DR3 mice days 10-52.
Scripts and functions for running multiple classification models (Feed forward neural network, linear discriminant analysis, support vector machine, naive bayes classifier) and exporting performance measures (classification accuracy with confidence intervals). Exhaustive search of parameters that maximize classification accuracy is performed and exported to excel file.
Note: parameters (variables) are first filtered on the basis of absolute value of log odds coefficient from elastic net regression (see other folders)
Analysis scripts for phenotypic classes derived from hierarchical clustering and heatmap representation.
Linear discriminant analysis per day (folder) of DR3 mice. Analysis considered combinations of 3 clusters for predicting class assignment and subsequent projection of results as a density distribution.
Note: parameters (variables) are first filtered on the basis of absolute value of log odds coefficient from elastic net regression (see other folders)
Elastic net (and lasso comparison) per day (10-52) of all cell types for predicting class assignment of mice. Waterfall plots included, along with cross validation summary plotting.
Cluster frequency to clinical measure correlation analysis with module detection, heatmap generation, and t-SNE creation. T-sne plots included that map elastic net clusters to t-SNE plot.
Scripts and data used to create circos plot