Galaxy |

lifelineskmcph (version 0.01)

Tabular input file for failure time testing.:

Must have a column with a measure of time and status (0,1) at observation.

Name of column containing a time to observation:

Use a column name from the file header if the data has one, or use one from the list supplied below, or use col1....colN otherwise to select the correct column

Status at observation. Typically 1=alive, 0=deceased for life-table observations:

Use a column name from the header if the file has one, or use one from the list supplied below, or use col1....colN otherwise to select the correct column

Optional comma delimited column names to use as covariates in the Cox Proportional Hazards model:

Leave blank for no Cox PH model tests

Title for this lifelines analysis:

Special characters will probably be escaped so do not use them

Optional comma delimited list of column names to use for this tabular file. Default is None when col1...coln will be used if no header row in the input data:

The column names supplied for time, status and so on MUST match either this supplied list, or if none, the original file header if it exists, or col1...coln as the default of last resort.

Optional group column name for KM plot:

If there are exactly 2 groups, a log-rank statistic will be generated as part of the Kaplan-Meier test.

Output format for all images:

This is a wrapper for some elementary life table analysis functions from the Lifelines package - see https://lifelines.readthedocs.io/en/latest for the full story

Given a Galaxy tabular dataset with suitable indicators for time and status at observation, this tool can perform some simple life-table analyses and produce some useful plots. Kaplan-Meier is the default. Cox Proportional Hazards model will be tested if covariates to include are provided.

Kaplan-Meier survival analysis - see https://lifelines.readthedocs.io/en/latest/Survival%20analysis%20with%20lifelines.html

This is always performed and a survival curve is plotted.

If there is an optional "group" column, the plot will show each group separately. If there are exactly two groups, a log-rank test for difference is performed and reported
The Cox Proportional Hazards model can be tested, if a comma separated list of covariate column names is supplied on the tool form.

These are used in as covariates.

Although not usually a real problem, some diagnostics and advice about the assumption of proportional hazards are are also provided as outputs - see https://lifelines.readthedocs.io/en/latest/Survival%20Regression.html

A big shout out to the lifelines authors - no R code needed - nice job, thanks!