survival_analysis

Function

crisgi_obj.survival_analysis(
    ref_group,
    target_group,
    interactions=None,
    groupbys=[],
    survival_types=['os', 'pfs'],
    time_unit='time',
    test_type='TER',
    method='prod',
    title=''
)

Performs survival analysis using Kaplan-Meier estimators and log-rank tests for specified groups and survival types. This method generates survival plots and statistical comparisons between groups, saving the resulting figures and printing output messages.

Parameters

Name	Type	Description
ref_group	str	Reference group name used for entropy calculation.
target_group	str	Comparison group name for analysis.
interactions	list or None	List of interaction features to include; if None, uses default from edata.uns.
groupbys	list	Additional columns in obs to group data by, in addition to score group.
survival_types	list	List of survival types to analyze (e.g., ['os', 'pfs']).
time_unit	str	Label for the time axis in plots (e.g., 'months', 'days').
test_type	str	Type of statistical test to use (default: 'TER').
method	str	Method for entropy calculation (default: 'prod').
title	str	Title prefix for the generated plots.

Return type

None

Returns

This function does not return a value. It generates and saves survival plots for each specified survival type and grouping, and prints the output file paths.

Attributes Set

• edata.obs['score']: Stores the computed score for each observation.
• edata.obs['score_group']: Stores the assigned score group for each observation.

Example

# Assume `crisgi` is an instance of the CRISGI class with loaded data

# Perform survival analysis comparing 'control' and 'treatment' groups
crisgi.survival_analysis(
    ref_group='control',
    target_group='treatment',
    interactions=['geneA', 'geneB'],
    groupbys=['batch'],
    survival_types=['os'],
    time_unit='months',
    test_type='DER',
    method='prod',
    title='Survival Analysis Example'
)

This example compares the overall survival ('os') between 'control' and 'treatment' groups, considering interactions for 'geneA' and 'geneB', and grouping by 'batch'. The resulting survival plot is saved to the output directory, and the file path is printed.