Skip to contents

LTFGRS is an R package that provides a unified interface to perform several phenotype-based family genetic risk scores (FGRS) as well as functions to perform the required data preparation and management steps.

The implemented methods are LT-FH++, PA-FGRS, and Kendler’s FGRS. The primary focus of the package is LT-FH++ and PA-FGRS. They share many similarities in their models, since they are both based on the same underlying liability threshold model with modifications. This package allows users to pick and choose which elements of those methods they want to use. Notable properties of the package are:

  • Automatic identification of family members up to n’th degree.
  • personalised thresholds for each included family that can account for:
    • age of diagnosis
    • censoring
    • cohort effects
    • sex
    • family-wise censoring
  • mixture distribution to handle partially observed controls.

Installation

You can install the development version of LTFGRS from GitHub with:

remotes::install_github("EmilMiP/LTFGRS")

You can install the LTFGRS from CRAN with:

install.packages("LTFGRS")

Documentation and Tutorials

Documentation and tutorials for the package can be found in the pkgdown site.

On the pkgdown site, a number of tutorial style vignettes can be found as well as smaller vignettes on data preparation and labelling of relatives. Please see the following tutorials:

Parallelisation in LTFGRS

LTFGRS bases its parallelisation on the future suite of packages. This means it is up to the user to decide when and how parallelisation should be applied. For details on how to use the future suite of packages, please see their documentation.

The function estimate_liability() is the function that has the highest computational burden and as such is supported by the future parallelisation framework.

A parallelisation setup example can be found in the vignette parallelisation guidance.