Multi-omics integration

• Integrate multiple omics layers: microbiome, metabolome, transcriptome, etc.
• Explore associations and predict outcomes
• Gain complementary insights that single-omics cannot provide

Data containers

• altExp() slot in TreeSummarizedExperiment
• MultiAssayExperiment

Methods

• Association: detect cross-omics correlations
• Ordination: reduce dimensionality, e.g., MOFA2, Joint-RPCA
• Supervised machine learning: predict outcomes, e.g., IntegratedLearner

Cross-association

• CLR + Spearman

Calculating “targeted” correlations

• Focus on feature pairs with found relationships
• Avoids computing all pairwise correlations
• Implemented in anansi (Bastiaanssen, Quinn, and Cryan 2023)

➡️ anansi slides

Ordination

Multi-omics Factor Analysis (MOFA2)

• Factorizes multi-omics matrix into latent factors
• Captures shared and data-specific variation across omics
• Outputs low-dimensional embeddings for samples and features

Ordination

Joint robust PCA (Joint-RPCA) (Cordazzo Vargas et al. 2026)

• Shared sample space across any number of omics, with feature loadings per omic
• Robust clr and matrix completion handle sparse, compositional data
• Cross-omic feature covariances for network analysis
• In mia: getJointRPCA(), addJointRPCA()

➡️ Joint-RPCA overview slides

Classical multi-table methods

• Mantel test, RV coefficient, co-inertia: are two tables associated?
• Sparse CCA: which features drive the association?
• STATIS, multiple factor analysis: combine more than two tables
• DPCoA: use the phylogenetic tree as prior knowledge

➡️ Multi-table methods slides

Supervised ML

IntegratedLearner

• Predict phenotypes from multiple omics
• Ensemble learning combining base learners per omics

➡️ IntegratedLearner slides

Demonstration

Exercises

From OMA online book, Chapter 23: Cross-association

• All exercises

References

Time series

A brief demonstration on time series tools.

• miaTime
• miaSim
• miaViz

MultiAssayExperiment example data

Load example data set and list the available experiments:

library(mia)
data(HintikkaXOData)
mae <- HintikkaXOData
experiments(mae)

ExperimentList class object of length 3:
 [1] microbiota: TreeSummarizedExperiment with 12706 rows and 40 columns
 [2] metabolites: TreeSummarizedExperiment with 38 rows and 40 columns
 [3] biomarkers: TreeSummarizedExperiment with 39 rows and 40 columns

MultiAssayExperiment example data

Each element in a MultiAssayExperiment is a (Tree)SummarizedExperiment.

experiments(mae)[["microbiota"]]

class: TreeSummarizedExperiment 
dim: 12706 40 
metadata(0):
assays(1): counts
rownames(12706): GAYR01026362.62.2014 CVJT01000011.50.2173 ...
  JRJTB:03787:02429 JRJTB:03787:02478
rowData names(7): Phylum Class ... Species OTU
colnames(40): C1 C2 ... C39 C40
colData names(0):
reducedDimNames(0):
mainExpName: NULL
altExpNames(0):
rowLinks: NULL
rowTree: NULL
colLinks: NULL
colTree: NULL

MultiAssayExperiment example data

Each element in a MultiAssayExperiment is a (Tree)SummarizedExperiment.

experiments(mae)[["metabolites"]]

class: TreeSummarizedExperiment 
dim: 38 40 
metadata(0):
assays(1): nmr
rownames(38): Butyrate Acetate ... Malonate 1,3-dihydroxyacetone
rowData names(0):
colnames(40): C1 C2 ... C39 C40
colData names(0):
reducedDimNames(0):
mainExpName: NULL
altExpNames(0):
rowLinks: NULL
rowTree: NULL
colLinks: NULL
colTree: NULL

MultiAssayExperiment example data

Each element in a MultiAssayExperiment is a (Tree)SummarizedExperiment.

experiments(mae)[["biomarkers"]]

class: TreeSummarizedExperiment 
dim: 39 40 
metadata(0):
assays(1): signals
rownames(39): Triglycerides_liver CLSs_epi ... NPY_serum Glycogen_liver
rowData names(0):
colnames(40): C1 C2 ... C39 C40
colData names(0):
reducedDimNames(0):
mainExpName: NULL
altExpNames(0):
rowLinks: NULL
rowTree: NULL
colLinks: NULL
colTree: NULL

Machine learning (Random Forests)

• You can always run any ML or other analysis method available in R by just extracting individual components from TreeSE (abundance assays and sample metadata)

• For important methods, the R/Bioconductor developer community tends to come up with more integrated solutions

• The development work always benefits from feedback