multiomics

Multiomics Integration Module

Core multi-omics data integration for combining GWAS, RNA-seq, proteomics, and epigenomics data layers.

Architecture

graph TD
    subgraph "Multiomics Integration"
        A[analysis/] --> |integration.py| I[Data Integration]
        M[methods/] --> |factorization.py| F[Matrix Factorization]
        M --> |clustering.py| CL[Multi-omic Clustering]
        P[pathways/] --> PA[Pathway Enrichment]
        S[survival/] --> SU[Survival Analysis]
        V[visualization/] --> VI[Integrated Visuals]
    end
    
    subgraph "Input Data"
        GWAS[GWAS Variants]
        RNA[RNA Expression]
        PROT[Protein Data]
    end
    
    GWAS --> I
    RNA --> I
    PROT --> I

Loading

Key Capabilities

GWAS-Expression (eQTL) Integration

The eQTL pipeline connects the GWAS Pipeline genomic variants with the Amalgkit RNA Pipeline expression abundances. See the eQTL Integration Pipeline for end-to-end execution.

You can also use the core functions directly to combine genetic variants with gene expression to identify regulatory mechanisms:

from metainformant.gwas.finemapping.colocalization import eqtl_coloc
from metainformant.multiomics.analysis import integration

# Prepare expression data for integration from a DataFrame or CSV/TSV matrix
rna_data = integration.from_rna_expression(expression_df, normalize=True)

# Prepare variant data from a VCF path or an existing sample-by-variant DataFrame
dna_data = integration.from_dna_variants("variants.vcf")

# Run colocalization analysis
result = eqtl_coloc(
    gwas_z=[1.2, 2.5, 3.1, 0.8],  # GWAS Z-scores
    eqtl_z=[1.1, 2.3, 2.9, 0.5],  # eQTL Z-scores
    gene_id="LOC12345"
)

Integration Methods

Method	Function	Use Case
integrate_omics_data()	Unified integration	Combine DNA/RNA/protein
joint_pca()	Joint dimensionality reduction	Find shared patterns
joint_nmf()	Non-negative factorization	Identify positive factors
canonical_correlation()	CCA	Correlate two data layers

Data Converters

Function	Purpose
from_dna_variants()	VCF path or variant matrix → integration-ready genotype dosage matrix
from_rna_expression()	Expression matrix loading, filtering, and normalization
from_protein_abundance()	Protein abundance matrix loading, filtering, and normalization
from_epigenome_data()	Methylation/ChIP-seq data

Submodules

Module	Purpose
analysis/	Core integration algorithms
methods/	Factorization, clustering
pathways/	Multi-omic pathway analysis
survival/	Cox, Kaplan-Meier models
visualization/	Integrated plots

Species-Agnostic Design

All methods are species-agnostic. Example with Apis mellifera:

# Works with any species - just provide the data
multiomics = integration.integrate_omics_data(
    dna_data=apis_variants_df,      # Any species VCF data
    rna_data=apis_expression_df,    # Any species expression
)

Related

• metainformant.gwas - GWAS analysis and fine-mapping
• metainformant.rna - RNA-seq quantification
• docs/multiomics/ - Extended documentation