Cross batch metabolite quantification harmonization

cross-batch-metabolite-quantification-harmonization

Summary

Filter and impute metabolites across multiple metabolomics batches to produce a unified feature set with complete quantification. This skill ensures only metabolites present across all batches are retained and missing values are resolved, enabling downstream batch-effect correction.

When to use

You have multiple SummarizedExperiment objects representing separate metabolomics assay batches with different metabolite coverage and missing-value patterns, and you need to identify a common metabolite set quantified across all batches before applying hierarchical batch normalization or comparative analysis.

When NOT to use

• Input is already a single unified feature table (not batch-separated SummarizedExperiment objects)
• You intend to retain batch-specific metabolites rather than enforce cross-batch consistency
• Missing-value mechanisms are informative and should not be imputed (e.g., absent metabolites indicate true biological absence rather than measurement failure)

Inputs

• List of SummarizedExperiment objects, each with log2-transformed metabolite abundance assay (logRaw or equivalent)
• Multiple batches of metabolomics data with metabolite × sample matrix structure

Outputs

• List of SummarizedExperiment objects with new assay (rawImpute) containing harmonized metabolites with imputed values
• Reduced metabolite set (e.g., 61 metabolites across 5 batches) quantified in all samples across all batches
• Assay matrix with no missing values

How to apply

Load metabolomics data as a list of SummarizedExperiment objects with log2-transformed assay (e.g., logRaw). Apply hRUV::clean() with threshold=0.5 to filter metabolites with >50% missing values per batch, method='intersect' to retain only metabolites quantified in all batches, and k-nearest neighbour imputation to fill remaining missing values. Extract and verify the output assay (e.g., rawImpute) for metabolite count and absence of missing data. The intersect method ensures strict cross-batch comparability by excluding any metabolite absent from even one batch.

Related tools

• hRUV (Primary tool: implements clean() function with threshold filtering, intersect method, and k-nearest neighbour imputation for cross-batch metabolite harmonization) — https://github.com/SydneyBioX/hRUV
• SummarizedExperiment (Data container: stores metabolite abundance matrices (assays) and metadata for batch-organized metabolomics data) — https://bioconductor.org/packages/SummarizedExperiment
• dplyr (Data manipulation and wrangling during workflow setup and verification)
• DMwR2 (Supports k-nearest neighbour imputation used by hRUV::clean())
• R (Execution environment)

Examples

dat_list = hRUV::clean(dat_list, threshold = 0.5, method = "intersect", assay = "logRaw", newAssay = "rawImpute")

Evaluation signals

• Output assay contains no missing values (NA count = 0) after imputation
• Metabolite count is reduced and consistent across all batches (e.g., 61 metabolites × n_samples for each batch)
• All retained metabolites are present in every batch (intersect condition verified)
• Metadata and sample names are preserved; only metabolite dimension is filtered
• rawImpute assay is correctly nested in each SummarizedExperiment object with proper assay naming and dimnames

Limitations

• Intersect method is stringent and may discard many batch-specific metabolites; not suitable when rare or batch-enriched metabolites carry biological significance
• k-nearest neighbour imputation assumes missing-at-random (MAR) mechanism; if missingness is informative (e.g., detection limits), imputation may bias downstream inference
• Threshold=0.5 (50% missing-value cutoff) is fixed in the demonstrated workflow; users must justify appropriateness for their data distribution
• Requires explicit batch structure in the SummarizedExperiment list; does not automatically detect or infer batch membership

Evidence

• [intro] threshold and intersect method ensure cross-batch metabolite retention: "filters metabolites with >50% missing values per batch using the intersect method, then performs k-nearest neighbour imputation"
• [intro] intersect method definition and strict cross-batch requirement: "selected metabolites that are quantified across all batches (intersect) with clean function"
• [intro] concrete hRUV::clean() function call with parameters: "dat_list = hRUV::clean(dat_list, threshold = 0.5, method = "intersect", assay = "logRaw", newAssay = "rawImpute")"
• [intro] input data structure requirement: "The data is already formatted in to a SummarizedExperiment object"
• [intro] expected output: metabolite reduction and missing-value elimination: "produce a rawImpute assay containing 61 metabolites quantified across all five batches"
• [readme] installation and dependency context: "Install the R package from GitHub using the devtools package"