The Polygenic Risk Score Knowledge Base offers a centralized online repository for calculating and contextualizing polygenic risk scores

Abstract

The process of identifying suitable genome-wide association (GWA) studies and formatting the data to calculate multiple polygenic risk scores on a single genome can be laborious. Here, we present a centralized polygenic risk score calculator currently containing over 250,000 genetic variant associations from the NHGRI-EBI GWAS Catalog for users to easily calculate sample-specific polygenic risk scores with comparable results to other available tools. Polygenic risk scores are calculated either online through the Polygenic Risk Score Knowledge Base (PRSKB; https://prs.byu.edu ) or via a command-line interface. We report study-specific polygenic risk scores across the UK Biobank, 1000 Genomes, and the Alzheimer's Disease Neuroimaging Initiative (ADNI), contextualize computed scores, and identify potentially confounding genetic risk factors in ADNI. We introduce a streamlined analysis tool and web interface to calculate and contextualize polygenic risk scores across various studies, which we anticipate will facilitate a wider adaptation of polygenic risk scores in future disease research.

Fig. 1. ADNI polygenic risk score distributions.

Alzheimer’s disease polygenic risk score distributions are shown for a ADNI participants with a CDR ≥ 1 compared to ADNI participants with a CDR ≤ 0.5 and b ADNI participants with a CDR ≥ 0.5 compared to ADNI participants with a CDR = 0.

Fig. 2. ADNI Polygenic Risk Scores using Lambert et al., 2013 GWA Summary Statistics.

PRSice-2 (dark grey), and the PRSKB (light grey) scores are shown. a PRSice-2 reports polygenic risk scores that center on 0, so 1.0 was added to each PRSice-2 score to put it on the same scale as the PRSKB, which centers polygenic risk scores based on odds ratios around 1.0. The PRSice-2 median score after transformation is 1.05207 and the PRSKB median score is 1.05338. b Since a polygenic risk score is a relative score compared to the sample population, we transformed the PRSKB scores by subtracting 0.00131 to overlap the shape of the distributions when both algorithms report the same median. Since the scores are normally distributed, a Welch’s two-sample t-test was used to determine the similarity between the two distributions, which were nearly identical (t = 0.004782; P = 0.9962).

Fig. 3. The PRSKB Tool Structure.

The PRSKB tool is composed of a client, a server, and a database. The user interacts with the client, which is either the web tool (https://prs.byu.edu), or the command-line interface (CLI). The client connects to the server that then retrieves and returns data from the PRSKB database to the client. The arrows in this diagram represent the flow of data. Boxes represent specific actions a PRSKB user can take with an icon indicating the client type for each box.

Fig. 4. Polygenic risk score workflow.

The process follows the standards established by Choi et al..