Skip to contents

PGS-TRI is for the analysis of polygenic scores (PGS) or polygenic risk scores (PRS) in case-parent trio studies for estimation of the risk of an index condition associated with direct effects of inherited PGS, indirect effects of parental PGS, and gene-environment interactions under a unified log-linear modeling framework. This tool can characterize genetic risks of diseases in the presence of population structure, assortative mating, and indirect genetic effects.

The full tutorial of PGS.TRI package and data examples are in this website https://ziqiaow.github.io/PGS.TRI/.

The source code for the manuscript is available in https://github.com/ziqiaow/PGS-TRI-Analysis.

Recent Version History

  • Apr 3, 2026: Added command line feature to run PGS-TRI.
  • Apr 2, 2026: Updated example codes.
  • March 30, 2026: Updated PGS-TRI function, added PGS-TRI-centered feature for centering δ\delta-IDE using reference data.
  • June 17, 2025: Updated citation medRxiv version 2.
  • Sept 18, 2024: Repository made public.

Run PGS-TRI in R

PGS-TRI can be run in an R session. An interactive R tutorial with examples is provided at: Get started.

Installation in R

The package can be easily installed from Github

library(devtools)
install_github("ziqiaow/PGS.TRI")

If you only want to use specific functions, you can also do this

rm(list = ls())
require(devtools)
source_url("https://github.com/ziqiaow/PGS.TRI/blob/main/R/PGS-TRI.R?raw=TRUE")
source_url("https://github.com/ziqiaow/PGS.TRI/blob/main/R/simulation.R?raw=TRUE")

#If directly downloaded the R files from Github to your local directory 
source("./R/PGS-TRI.R")
source("./R/simulation.R")

Run PGS-TRI from Command Line

PGS-TRI can also be run directly from the command line without starting an R session. Below are example codes for running PGS-TRI using command line.

Getting Started 

# Download PGS.TRI package
git clone https://github.com/ziqiaow/PGS.TRI.git

# Go to the directory
cd PGS.TRI

# Load R (if not already loaded)
module load R

Test with example data

Below is a toy example. Replace path_to_package, and path_to_result with your own directory paths. The example data is saved in the repository ~/PGS.TRI/inst/extdata/testdat_snipar.txt.

path_to_package='/dcs04/nilanjan/data/zwang/tools'
path_to_result='/dcs04/nilanjan/data/zwang/tools/PGS.TRI/inst/extdata'

Rscript ${path_to_package}/PGS.TRI/inst/pgs_tri \
  --input ${path_to_package}/PGS.TRI/inst/extdata/testdat_snipar.txt \
  --output ${path_to_result}/test_results.txt \
  --offspring-col pgs_c \
  --mother-col pgs_m \
  --father-col pgs_f \
  --save-rds \
  --verbose

If we want to test for indirect effect, and also center by a reference data (such as the UK Biobank) female-male PGS value difference of 0.0001:

Rscript ${path_to_package}/PGS.TRI/inst/pgs_tri \
  --input ${path_to_package}/PGS.TRI/inst/extdata/testdat_snipar.txt \
  --output ${path_to_result}/test_results.txt \
  --offspring-col pgs_c \
  --mother-col pgs_m \
  --father-col pgs_f \
  --parental-indirect TRUE \
  --parental-diff-ref 0.0001 \
  --save-rds \
  --verbose

The result output (.txt output and .rds file) will be saved in path_to_result.

Basic analysis

Replace path_to_package, path_to_data, and path_to_result with your own directory paths. Replace data.txt and results.txt with your own data and output names.

Rscript ${path_to_package}/PGS.TRI/inst/pgs_tri --input ${path_to_data}/data.txt --output ${path_to_result}/results.txt

Full model

Specify the column names from your input .txt file for child (–offspring-col), mother (–mother-col), and father (–father-col), for GxE interaction analysis, the column names for E variables should be specified in –envir and –formula.

Rscript ${path_to_package}/inst/pgs_tri \
  --input ${path_to_data}/data.txt \
  --output ${path_to_result}/results.txt \
  --offspring-col pgs_c \
  --mother-col pgs_m \
  --father-col pgs_f \
  --parental-indirect \
  --parental-diff-ref 0 \
  --gxe \
  --envir age,sex,bmi \
  --formula '~ age + factor(sex) + bmi' \
  --small-trio \
  --save-rds \
  --verbose

Command Line Options 

Required:
  --input FILE              Input file, see format below. Each row presents a family, each column represents the PGS values of child, mother, father, respectively; if GxE interactions analysis is required, then environmental variables can also be included in the same .txt file
  --output FILE             Output file path

Analysis Options:
  --gxe                     Estimate gene-environment interactions [default: FALSE]
  --parental-indirect       Estimate parental indirect effects [default: FALSE]
  --parental-diff-ref NUM   Reference data female-male PGS difference [default: 0]
  --envir VARS              Environmental variables (comma-separated)
  --formula FORMULA         R formula for environmental variables
  --side INT                Test side: 1 or 2 [default: 2]
  --small-trio              Use t-test for small samples (N_trio < 100) [default: FALSE]

Input Format:
  --offspring-col NAME      Offspring PGS column [default: pgs_c]
  --mother-col NAME         Mother PGS column [default: pgs_m]
  --father-col NAME         Father PGS column [default: pgs_f]

Output Options:
  --save-rds                Save results as RDS files [default: TRUE]
  --verbose                 Print detailed output [default: TRUE]
  --help                    Show this help message

Input File Format

Tab-delimited file with family PGS data:

pgs_c   pgs_m   pgs_f   age sex
0.523   0.412   0.389   45  1
0.678   0.534   0.621   52  0
0.234   0.289   0.198   38  1

For GxE analysis, include environmental variables as additional columns.

Questions

This software will be constantly updated, so please send your questions/suggestions to to help improve the package.

Reference

  • Ziqiao Wang, Luke Grosvenor, Debashree Ray, Ingo Ruczinski, Terri Beaty, Heather Volk, Christine Ladd-Acosta, Nilanjan Chatterjee (2024). Estimation of Direct and Indirect Polygenic Effects and Gene-Environment Interactions using Polygenic Scores in Case-Parent Trio Studies. medRxiv. https://www.medrxiv.org/content/10.1101/2024.10.08.24315066v2