Multi-Threshold Survival Analysis

Implements survival analyses across multiple abundance thresholds, repeatedly partitioning samples into groups and evaluating survival differences to assess taxonomic associations with outcomes.

mt.surv

Multi-threshold survival analysis

Overview

This package contains functions to apply multi-threshold survival analysis on count data (e.g. microbe abundances, gene expression), using (1) survival information and (2) counts.

survivalByQuantile generates a table with cox analysis parameters at various threshold quantile

calculateArea integrates the area above a 0.05 log likehood p-value threshold and outputs a scalar that can be used to rank the importance of different features

ToWide converts the count matrix reads from long format to wide format

Installation

# Currently available as the development version on GitHub
install.packages("devtools")
devtools::install_github("spakowiczlab/mt.surv")

References

1. Andersen, P. and Gill, R. (1982). Cox's regression model for counting processes, a large sample study. Annals of Statistics 10, 1100-1120.

2. Therneau, T., Grambsch, P., Modeling Survival Data: Extending the Cox Model. Springer-Verlag, 2000.

submitted