# ApplyPolygenicScore: An R package for applying polygenic risk score models

## TL;DR

ApplyPolygenicScore is an R package that helps researchers apply genetic risk models to predict complex traits like BMI in cancer patients.

## Contribution

The package introduces a streamlined, automated tool for applying polygenic risk models to new genetic datasets.

## Key findings

- ApplyPolygenicScore includes functions for allele matching and PGS computation with visualization.
- A BMI PGS applied to cancer patients showed limited accuracy, highlighting non-genetic factors.
- The package is open-source and well-documented for broader research use.

## Abstract

A polygenic score (PGS) predicts an individual’s genetic predisposition to a complex trait. A PGS is created by estimating the relative contributions of multiple common variants to the overall trait, creating a polygenic risk model (PGM). The PGM is then applied by combining its weights with the genotypes of a specific individual to estimate individual-specific genetic predisposition. Genome-wide association studies have served as the basis for thousands of PGMs, leading to many studies associating PGSs with a range of outcomes.

To simplify, improve, and automate this task, we developed ApplyPolygenicScore, an open-source R package for applying standardized PGMs to new genetic data. We demonstrate its capabilities in a case study, applying a PGM for body mass index (BMI) in 1071 patients diagnosed with bladder, liver, and endometrial cancer.

ApplyPolygenicScore includes functions for input validation, allele matching, and PGS computation and visualization and is extensively documented. The computed PGS for BMI predicted BMI in patients with cancer, but its low accuracy indicates a larger role for nongenetic factors in BMI-influenced cancer outcomes.

ApplyPolygenicScore encourages the wider research community to extend the findings of the statistical genetics niche, facilitating broader use of PGSs and subsequent novel discovery.