In maialab/quincunx: REST API Client for the 'PGS' Catalog

knitr::opts_chunk$set(
  collapse = TRUE,
  comment = "#>"
)
library(quincunx)

Overview of polygenic risk score construction

PGS is calculated as a weighted sum of several risk variants from a genome-wide association study in one or more samples with multiple p-value thresholds. The effect sizes are typically estimated as $\beta$ (beta coefficients) or as odds ratios. After the PGS is calculated in one sample, the distribution of individual PGS is assessed in another in an independent sample set.

knitr::include_graphics('../man/figures/pgs-construction-process.png') 

PGS development and evaluation stages

knitr::include_graphics('../man/figures/pgs-stages.svg') 

In the PGS Catalog, cohorts and samples are annotated according to their utilisation context, i.e. stage, in the PGS construction process. In quincunx, the stage is indicated by the stage variable that can have one of these values:

• gwas: to annotate samples used to derive variant associations (GWAS)
• dev: to annotate samples used in the development or training of PGSs
• gwas/dev: as a catch-all term to annotate samples used either in gwas or dev stages
• eval: to annotate samples used in the PGS evaluation stage

You will encounter the stage annotation in tables of objects returned by quincunx's retrieval functions. Here are a few examples:

In a scores object

get_scores('PGS000327')@samples

In a sample_sets object

get_sample_sets(pgs_id = 'PGS000327')@samples

In a performance_metrics object

get_performance_metrics(pgs_id = 'PGS000327')@samples

In the stages_tally table:

get_scores('PGS000327')@stages_tally

In the ancestry_frequencies table:

get_scores('PGS000012')@ancestry_frequencies

And in multi_ancestry_composition table:

get_scores('PGS000012')@multi_ancestry_composition

In a cohorts object:

get_cohorts('23andMe')@pgs_ids

maialab/quincunx documentation built on Aug. 18, 2022, 5:31 a.m.