imputomics 2.0

Extending imputomics 1.0 from metabolomics to multi-omics: a standardized wrapper around state-of-the-art MVI algorithms, a curated repository, and a rigorous benchmarking suite

🧠 Missing values distort multi-omics analyses

imputomics 2.0: comprehensive framework for missing value imputation in multi-omics data

PI: Jarek
Funding agency: National Science Center Poland
Programme: Miniatura 9
Grant ID: 2025/09/X/NZ7/00157
Start date: 25 Jun 2025
Duration: 12 months
Budget: PLN49,999.00

Missing values are pervasive in omics (genomics, transcriptomics, proteomics, metabolomics) and can compromise downstream analyses. Missing-value imputation algorithms (MVIAs) attempt to replace missing entries with plausible estimates. Given the diversity of data types and the sheer number of methods, choosing the right MVIA remains challenging.

Building on imputomics 1.0 (web server + R package) — where we conducted an extensive preliminary benchmark, imputomics 2.0 generalizes the framework to multi-omics, integrates state-of-the-art MVIAs, and provides a curated repository of methods and datasets.

📌 Project highlights

  • 📚 Comprehensive review of missing value imputation algorithms (MVIAs)
  • 🗂️ Standardized annotation of omics-specific imputation methods
  • ⚙️ Universal wrapper for seamless integration across genomics, proteomics, metabolomics & more
  • 📊 Benchmarking with multiple metrics (RMSE, correlation, interpretability, downstream tasks)
  • 🌐 Building an open repository of MVIAs + datasets for the research community