Skip to contents

Introduction to pleior

Source: vignettes/pleior.Rmd
pleior.Rmd

Abstract

Pleiotropy occurs when a single genetic variant influences multiple traits, offering insights into shared biological pathways and disease relationships. The package provides tools for analyzing pleiotropy in genome-wide association studies (GWAS).

This vignette demonstrates a typical workflow using GWAS catalog data to identify and visualize pleiotropic genetic variants.

What is Pleiotropy?

Pleiotropy is a fundamental concept in genetics where a single gene or genetic variant affects multiple phenotypic traits. In GWAS studies:

  • Genetic Pleiotropy: One variant associated with multiple traits
  • Biological Significance: Reveals shared pathways and mechanisms
  • Clinical Relevance: Explains disease comorbidity and drug targets

Understanding pleiotropy helps the users to:

  • Identify shared genetic architecture across diseases
  • Discover novel therapeutic targets
  • Explain why certain diseases co-occur
  • Prioritize variants for functional studies

Installation 

# Install from GitHub
if (!require("devtools", quietly = TRUE))
    install.packages("devtools")

devtools::install_github("danymukesha/pleior")

# Load the package
library(pleior)

Quick Start with Real GWAS Data

The package includes a curated subset of GWAS catalog data containing ~2,500 significant associations across 15 diverse traits.

# Load the example dataset
data(gwas_subset)

# Explore the data
head(gwas_subset)

# Summary statistics
cat("Dataset contains:", nrow(gwas_subset), "GWAS associations\n")
cat("Unique SNPs:", length(unique(gwas_subset$SNPS)), "\n")
cat("Unique traits:", length(unique(gwas_subset$MAPPED_TRAIT)), "\n")

Complete Pleiotropy Analysis Workflow

Step 1: Preprocess GWAS Data

Clean the data and filter for genome-wide significance:

# Filter for genome-wide significant associations (p < 5e-8)
gwas_clean <- preprocess_gwas(
    gwas_subset,
    pvalue_threshold = 5e-8
)

cat("After filtering:", nrow(gwas_clean), "significant associations\n")

Step 2: Detect Pleiotropic SNPs

Identify SNPs associated with multiple traits:

# Detect pleiotropy across all traits
pleio_results <- detect_pleiotropy(gwas_clean)

cat("Found", nrow(pleio_results), "pleiotropic SNPs\n")

# Examine the top pleiotropic SNPs
head(pleio_results, 10)

Step 3: Analyze Pleiotropic SNPs 

# Summary statistics
snp_summary <- pleio_results %>%
    group_by(SNPS) %>%
    summarise(
        n_traits = n(),
        min_pvalue = max(PVALUE_MLOG),
        .groups = "drop"
    ) %>%
    arrange(desc(n_traits))

print(snp_summary)

Step 4: Visualize Results

Manhattan Plot

Create a Manhattan plot highlighting pleiotropic SNPs:

# Plot with top pleiotropic SNP highlighted
if (nrow(pleio_results) > 0) {
    top_snp <- snp_summary$SNPS[1]

    p_manhattan <- plot_pleiotropy_manhattan(
        pleio_results,
        highlight_snp = top_snp,
        title = "Pleiotropic SNPs Across Genome"
    )

    print(p_manhattan)
}

Trait Distribution

See how many SNPs are associated with each trait:

trait_counts <- pleio_results %>%
    count(MAPPED_TRAIT, sort = TRUE)

# Simple bar plot
library(ggplot2)

p_traits <- ggplot(trait_counts, aes(x = reorder(MAPPED_TRAIT, n), y = n)) +
    geom_col(fill = "steelblue") +
    coord_flip() +
    labs(
        x = NULL,
        y = "Number of Pleiotropic SNPs",
        title = "Pleiotropic SNPs by Trait"
    ) +
    theme_minimal()

print(p_traits)

Pleiotropy Spectrum

Visualize the distribution of number of traits per SNP:

snp_trait_dist <- pleio_results %>%
    group_by(SNPS) %>%
    summarise(n_traits = n(), .groups = "drop")

p_spectrum <- ggplot(snp_trait_dist, aes(x = n_traits)) +
    geom_histogram(
        binwidth = 1,
        fill = "steelblue",
        color = "white",
        alpha = 0.8
    ) +
    labs(
        x = "Number of Traits per SNP",
        y = "Count",
        title = "Distribution of Pleiotropy"
    ) +
    theme_minimal()

print(p_spectrum)

Advanced: Focus on Specific Traits

You can analyze pleiotropy between specific traits of interest:

# Define traits of interest
target_traits <- c(
    "Alzheimer's disease",
    "Type 2 diabetes",
    "Coronary artery disease"
)

# Filter for these traits
gwas_target <- gwas_clean %>%
    filter(MAPPED_TRAIT %in% target_traits)

# Detect pleiotropy among target traits
pleio_target <- detect_pleiotropy(
    gwas_target,
    traits = target_traits
)

cat("Pleiotropy between", length(target_traits), "target traits:\n")
print(pleio_target)

Interpreting Results

What Makes a SNP “Pleiotropic”?

A SNP is considered pleiotropic if it shows significant associations with two or more distinct traits in our dataset. The package identifies:

  1. SNP Identifier: The variant name (e.g., rs814573)
  2. Associated Traits: Which traits the SNP influences
  3. Significance Level: The strongest association (-log10 p-value)
  4. Genomic Location: Chromosome and position

Example Interpretation

Consider a SNP like rs814573:

  • Associated with: Alzheimer’s disease, myocardial infarction, LDL cholesterol
  • Significance: p < 5e-8 for all traits
  • Implication: This variant affects pathways related to lipid metabolism and cardiovascular health
  • Research Direction: Study shared biological mechanisms between these diseases

Next Steps

After identifying pleiotropic SNPs:

  1. Functional Annotation: Map SNPs to genes and pathways
  2. Literature Review: Investigate known mechanisms
  3. Experimental Validation: Test in model systems
  4. Drug Target Assessment: Evaluate therapeutic potential

For advanced visualization methods including heatmaps, networks, and publication-ready themes, see the Visualization for Pleiotropy Analysis vignette.

Performance Tips

For large GWAS datasets:

# Use data.table for efficient operations
library(data.table)

# Parallel processing can speed up analysis
library(future)
plan(multisession)

# Cache preprocessed results
if (file.exists("preprocessed_gwas.rds")) {
    gwas_clean <- readRDS("preprocessed_gwas.rds")
}

Getting Help 

# Function documentation
?detect_pleiotropy
?preprocess_gwas
?plot_pleiotropy_manhattan

# Package help
help(package = "pleior")

References

Session Information