Thanks to visit codestin.com
Credit goes to github.com

Skip to content

fubar2/lifelines_tool

Repository files navigation

lifelines_tool - lifelines survival analysis python package wrapped as a Galaxy tool.

  1. A Galaxy tool for right censored failure time data.
  2. Provides Kaplan-Meier plots with confidence intervals, and optional Cox proportional hazards models
  3. Uses the lifelines package.
  4. Install to your Galaxy server from the toolshed - search for lifelines_km_cph_tool owned by fubar2

Using the Rossi sample recidivism data from lifelines:

Runs Kaplan-Meier and generates a plot. Optional grouping variable race was supplied, so plots show confidence intervals

KM plot sample

If 2 groups, runs a log-rank test for difference in the report.

KM plot sample

If a comma separated list (for example: prio, age, race, mar, fin) of covariate column names is provided, a Cox proportional hazards model is run, the assumption of proportionality is tested, and recommendations made.

KM plot sample

Partial plots for each covariate. Quintiles are used if there are >10 distinct values in the column. If you have non-ordinal covariates with >10 categores, this will produce garbage plots, but the model will also be hard to interpret so good luck.

C-PH partial plot samples

C-PH partial plot samples

Uses pandas read_csv with tabular delimiters so should work with any tabular data with the required columns - time and status for observations.

Issues to https://github.com/fubar2/lifelines_tool please. Autogenerated so pull requests are possibly meaningless but regeneration of a new version should work.

Tool made with the Galaxy ToolFactory: https://github.com/fubar2/galaxy_tf_overlay

Galaxy Training Network tutorials are available here

The generating history for this tool can be imported into a ToolFactory instance using this url: https://github.com/fubar2/lifelines_tool/raw/main/Galaxy-ToolFactory-History-lifelineskmcphtoolfactory.tar.gz

Once imported, switch to that new history. The original ToolFactory form can be recreated using the redo button. Editing the tool id will make a new tool, so all other edits to parameters can be made and the new tool generated without destroying the original sample.

About

Wraps lifelines for KM and CPH models from tabular data

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages