MR Data Generator

Step 1/3: Data Upload & Harmonization

Upload CSV, tab-delimited TEXT, JSON, or GZ files containing exposure and/or outcome data. Files are automatically harmonized upon processing.

Upload Exposure Data File

Expected CSV format:

SNP,ea,nea,beta,se,eaf
rs9939609,A,T,0.15,0.02,0.45
rs17782313,C,T,0.08,0.015,0.32
rs71358633,G,A,0.12,0.018,0.28

Expected JSON format:

{"rs9939609": {"ea":"A","nea":"T","beta":0.15,"se":0.02,"eaf":0.45},"rs17782313": 
{"ea":"C","nea":"T","beta":0.08,"se":0.015,"eaf":0.32}}

Upload Outcome Data File

Expected CSV format:

SNP,beta,se,ea,nea,eaf
rs9939609,-0.08,0.03,A,T,0.45
rs1801282,0.10,0.025,C,G,0.35
rs662799,-
0.05,0.012,A,G,0.78

Expected JSON format:

{"rs9939609": {"beta":-0.08,"se":0.03,"ea":"A","nea":"T","eaf":0.45},"rs1801282": 
{"beta":0.10,"se":0.025,"ea":"C","nea":"G","eaf":0.35}}

Upload files to begin

Step 2/3: Correlation Matrix

Download correlation matrix from genetic databases for Mendelian Randomization analysis. Select your preferred database source below. If database access fails, a blank correlation matrix will be automatically generated.

⬇️ Download Correlation Matrix

Download pre-computed correlation matrix from genetic databases (1000 Genomes, UK Biobank, gnomAD).

Note: For UK Biobank and gnomAD, you may need API tokens. 1000 Genomes is publicly accessible. Blank matrices are automatically created if database access fails.

Source Database

Status: Ready to download correlation matrix. Blank matrix will be created if database access fails.

Correlation Matrix Status: Download required to proceed to MR analysis.

Ready to download correlation matrix (blank matrix available if needed)

Step 3/3: MR Input Construction

Generate MR input objects with integrated correlation matrix for advanced analysis methods (IVW-MRE, MR-Egger, Weighted Median, etc.). Blank correlation matrices are automatically handled.

🔗 Correlation Matrix Integration

Status: No correlation matrix available

Univariate MR Input (with Correlation)

Generate single-exposure MR input with correlation matrix for IVW-MRE, MR-Egger, and other methods.

Real MR input with correlation 
from genetic databases will appear here:
{
  "snps": ["rs9939609", "rs17782313", "rs71358633"],
  "betaX": [0.15, 0.08, 0.12],
  "betaY": [-0.08, 0.10, -0.05],
  "betaXse": [0.02, 0.015, 0.018],
  "betaYse": [0.03, 0.025, 0.012],
  "correlation": [[1.0, 0.45, 0.23], [0.45, 1.0, 0.38], [0.23, 0.38, 1.0]],
  "exposure": "BMI",
  "outcome": "Type2Diabetes",
  "n": 3,
  "correlation_info": {
    "source": "1000genomes",
    "population": "EUR",
    "method": "precomputed"
  }
}

Exposure Name Outcome Name

Multivariate MR Input (with Correlation)

Generate multi-exposure MR input with correlation matrix for multivariable analysis.

Real multivariate MR input 
with correlation from genetic databases will appear here:
{
  "snps": ["rs9939609", "rs17782313", "rs71358633"],
  "betaX": [[0.15, 0.08], [0.08, 0.12], [0.12, 0.10]],
  "betaY": [-0.08, 0.10, -0.05],
  "betaXse": [[0.02, 0.015], [0.015, 0.018], [0.018, 0.025]],
  "betaYse": [0.03, 0.025, 0.012],
  "correlation": [[1.0, 0.45, 0.23], [0.45, 1.0, 0.38], [0.23, 0.38, 1.0]],
  "exposure": ["BMI", "CRP"],
  "outcome": "Type2Diabetes",
  "n": 3,
  "correlation_info": {
    "source": "1000genomes",
    "population": "EUR",
    "method": "precomputed"
  }
}

Exposure Names (comma-separated) Outcome Name

Complete Analysis Export

Export Preview

Complete analysis package will appear here with all MR inputs and 
correlation matrix integrated from genetic databases.

Ready to generate MR inputs

Unit Tests (Browser-based)

Pending

Example test output:
✔ File parsing - normalizeRow (rs9939609, rs17782313 validated)
✔ Auto-harmonization alignment (rs9939609 alleles A/T matched)  
✔ Correlation matrix generation - Database API integration successful
✔ MR input with correlation - 3 SNPs processed with real correlation matrix
✔ Export validation - Database correlation info included in all outputs
Completed: 5/5 passed