GWAS / Molecular QTL Visualization Platform

📖 Platform Overview

The GWAS Visualization Platform is a professional tool for Genome-Wide Association Study visualization, providing fast rendering of Manhattan plots, QQ plots, cis-eQTL plots, and trans-eQTL plots.

This platform uses intelligent rasterization technology to accelerate plotting of large datasets by up to 20x , and supports high-resolution PNG/PDF output for publication-quality figures.

🚀 Quick Start

1️⃣ Data Upload

Supported file formats:

• .txt - Tab-delimited text file
• .csv - Comma-separated text file
• .tsv - Tab-separated text file
• .mlma - GCTA software output format

File size limit: max 1 GB

Data requirements: must contain at least chromosome, position, and P-value columns (for Manhattan/QQ plots)

2️⃣ Auto Column Detection

The system automatically recognizes the following common column names (case-insensitive):

• Chromosome column: chr, chromosome, chrom, snp_chr, chr_name
• Position column: pos, position, bp, ps, snp_pos, base_pair
• P-value column: p, pval, pvalue, p_value, pv, p_bolt_lmm
• Gene column: gene, gene_name, geneid, symbol

Note: Auto-detection accuracy is ~90%. If incorrect, adjust columns manually.

3️⃣ Generate Plot

Select the corresponding plot type:

• 📊 Manhattan & QQ: Standard GWAS visualization, displayed side by side
• 🧬 cis-eQTL: Manhattan plot with gene annotation, labeling only the most significant SNP per gene
• 🔀 trans-eQTL: Distal association visualization with 1:1 square layout

Click 🎨 Generate Plot and wait for the plot to render

Click ⬇️ Download Image to save the high-resolution figure locally

⚙️ Feature Details

Manhattan Plot

Standard GWAS visualization showing significance levels for all SNPs:

• Color schemes: 6 options (Classic, Modern, Colorblind-friendly, Pastel, Vibrant, Grayscale)
• Significance threshold: Default 5e-8 (genome-wide significance), customizable
• Rasterization: Auto-enabled when data exceeds 100k points, 20x speedup

QQ Plot

Evaluates whether the P-value distribution matches expectations; detects population stratification and inflation:

• Red dashed line represents the expected distribution
• Points deviating from the line indicate inflation or true signals
• 16x speedup when data exceeds 100k points

cis-QTL Manhattan

cis-QTL visualization with intelligent gene annotation:

• Gene annotation threshold: Uses -log10(P) units, consistent with Y-axis (e.g., 6 means P < 1e-6)
• Smart deduplication: Only the most significant SNP per gene is labeled
• ggrepel annotation: Automatically avoids label overlap with connecting lines
• Optional threshold line: Toggle the red significance threshold line on/off

trans-QTL Plot

trans-QTL visualization showing distal regulatory relationships:

• 1:1 layout: Same scale on both axes, square figure
• Chromosome order: Both axes strictly ordered 1, 2, 3...
• Effect-size coloring: Cyan (negative) - White (zero) - Pink (positive)
• Smart rasterization: 9x speedup for large datasets

🎛️ Parameter Reference

Column Position

Specify the position of each column in the data file (1-indexed):

Significance Threshold

• Manhattan plot: uses P-value units (e.g., 5e-8), controls the red threshold line position
• cis-QTL gene annotation: uses -log10(P) units (e.g., 6), controls which genes are labeled

Output Settings

• Format: PNG (for preview) or PDF (for publication, vector)
• Size: Width and height in inches
• DPI: Resolution — 300 for publication, 600 for high-res printing

⚡ Performance Optimization

Intelligent Rasterization

When data exceeds 100k points, rasterization is automatically enabled:

• Manhattan: from 90s down to 4.5s (20x speedup)
• QQ Plot: from 60s down to 3.8s (16x speedup)
• trans-QTL: from 200s down to 22s (9x speedup)

Note: Rasterization does not affect plot quality; visual appearance is fully preserved.

Data Processing Tips

• < 100k points: no special handling needed, upload directly
• 100k - 1M points: recommend enabling 'Rasterization'
• > 1M points: rasterization is applied automatically; rendering takes ~5-30 seconds

❓ Frequently Asked Questions

Q1: Why are my chromosomes in the wrong order?

A: Check that the chromosome column is in numeric format. If it is character format (e.g., chr1, chr2), the system sorts alphabetically. Use plain integers (1, 2, 3...) for correct ordering.

Q2: Too many gene labels — how do I reduce them?

A: Increase the 'Gene Annotation Threshold' (-log10P), e.g., from 6 to 8 or 10, to annotate only more significant genes. The system annotates at most 50 genes by default.

Q3: Why is the trans-QTL plot a square?

A: To make it easier to observe the diagonal (cis associations) and off-diagonal (trans associations), a 1:1 aspect ratio is used.

Q4: How do I get the highest-quality output?

A: Use the following settings:

• Format: PDF (vector, losslessly scalable)
• DPI: 600 (ultra-high resolution)
• Size: adjust according to journal requirements

Q5: Which GWAS software outputs are supported?

A: Most GWAS software is supported, including:

• PLINK (.assoc, .qassoc)
• GCTA (.mlma)
• BOLT-LMM (.stats)
• SAIGE (.txt)
• REGENIE (.regenie)
• Any other tabular file containing chromosome, position, and P-value columns

📌 Version Information

Current Version: v1.3.2

Release Date: 2026-01-14

Key Features:

• ✅ Intelligent rasterization (20x speedup)
• ✅ Auto column detection (90% accuracy)
• ✅ Manhattan & QQ side-by-side display
• ✅ cis-eQTL gene annotation (ggrepel)
• ✅ trans-eQTL 1:1 square layout
• ✅ Chromosomes in ascending order
• ✅ Data preview (all plot pages)

Tech Stack:

• Shiny: Web application framework
• ggplot2: Plotting engine
• data.table: High-performance data processing
• ggrepel: Intelligent label layout

💬 Need Help?

If you have questions or suggestions, feel free to reach out:

Email | GitHub | FAQ