inst/markdown/cox_regression.markdown

In CRI-iAtlas/iatlas-app: What the Package Does (One Line, Title Case)

Title: Cox Proportional Hazard Ratio Models

1. Univariable Cox Proportional Hazard Model:

In this case, a Cox Proportional Hazard model for each combination of dataset and feature is created, using the coxph function of the R package survival (Thermeau et al., 2020).

To correct for multiple testing, Benjamini-Hochberg method is computed on p-values from all the models of the same dataset. This step uses the p.adjust function from the R package stats (R Core Team, 2019).

In this analysis, the Hazard Ratio coefficients are independent, ie., the computed values are not changed if you add or delete features.

Example

• User Selection:

  Survival endpoint: Overall Survival (OS)

  Datasets selected: Gide 2019, Van Allen 2015

  Features selected: Wound Healing, IFN-gamma signature

• Models created:

For samples from Gide 2019:

Model1. OS ~ Wound Healing

Model2. OS ~ IFN-gamma signature

For samples from Van Allen 2015:

Model3. OS ~ Wound Healing

Model4. OS ~ IFN-gamma signature

• Benjamini-Hochberg False Discovery Rate correction:

  p-values from Model1 and Model2 -> p.adjust()

  p-values from Model3 and Model4 -> p.adjust()

2. Multivariable Cox Proportional Hazard Model:

In this option, it is computed, for each dataset, a Cox Proportional Hazard model of time to survival outcome considering all the selected features as predictors. As a consequence, the inclusion or exclusion of features can change the computed values for coefficients in a model. This analysis also uses the coxph function of the R package survival (Thermeau et al., 2020).

Example

• User Selection:

  Survival endpoint: Overall Survival (OS)

  Datasets selected: Gide 2019, Van Allen 2015

  Features selected: Wound Healing, IFN-gamma signature

• Models created:

  For samples from Gide 2019: OS ~ Wound Healing + IFN-gamma signature

  For samples from Van Allen 2015: OS ~ Wound Healing + IFN-gamma signature

References:

R Core Team (2019). R: A language and environment for statistical computing. R Foundation for Statistical Computing, Vienna, Austria.

Therneau T (2020). A Package for Survival Analysis in R. R package version 3.1-12, .

CRI-iAtlas/iatlas-app documentation built on Feb. 7, 2025, 9:02 p.m.