groupedSurv

groupedSurv implements grouped survival modeling in R for grouped failure time phenotypes using an exact likelihood approach to enable genome-wide analyses of imprecise or interval-recorded event data such as drug-induced toxicity studies.

Key Features:

• Exact Likelihood Function: Employs an exact likelihood function to estimate grouped survival models and account for data grouping during inference.
• Handling Grouped Failure Time Data: Models event times or doses recorded within discrete intervals rather than as exact values.
• Adjustments for Covariates: Supports inclusion of baseline covariates to adjust for confounding variables.
• Genome-Wide Analysis Capability: Performs analyses at variant, gene, or pathway levels for genome-wide genetic association studies.
• Computational Efficiency: Designed for computational efficiency with large genome-wide datasets.

Scientific Applications:

• Genetic Association Studies: Reanalyzes published genome-wide studies to identify common germline variants associated with drug-induced toxicities, exemplified by studies of taxane-induced peripheral neuropathy in breast cancer patients.
• Translational Research: Applicable to translational studies where precise event timing is challenging due to experimental or data collection limitations.

Methodology:

Estimation is based on an exact likelihood function for grouped survival models, and the method was validated by simulating statistical properties and computational performance.