ROADMAP 2025

[kuwisdelu]

This issue serves as a tentative outline of development planned for 2025.

IF YOU ARE A USER, DO NOT MAKE RESEARCH PLANS BASED ON THIS TIMELINE.

[kuwisdelu]

Spring 2025

Cardinal 3.10 (Bioconductor 3.21 - expected April)

This will be our 10th anniversary edition! We can be ambitious. Among other new features, this should finally be the release where we introduce a GUI.

• Graphical user interface
  • Based on shiny
  • Visualization only (for now)
  • Maybe preview pre-processing?
  • iMSI()? MSIView()? msiView()?
• Ion mobility (experimental)
  • Support for timsTOF via TIMSCONVERT to imzML
  • Add readTIMSData() function
  • Read mobility arrays with cvParam "MS:1003006"
  • Load data as TIMSImagingExperiment
  • Add peakProcess() method for TIMSImagingExperiment
  • Pre-release on Github in February/March?
• Co-registration (experimental)
  • Native support for simple linear and affine transformations
  • Add coregister method based on matter::warp2_trans()
  • Aim to be good enough for visualization, but not necessarily annotation
  • More advanced methods coming later in CardinalNN

CardinalNN 0.99 (under development)

This will be our companion package for any neural network-based methods requiring Python.

• Prototype R package
• Define Basilisk virtual environments
• Methods for serializing core objects to Python
  • Need to figure out how to map to Python classes
  • May need to pull data into memory for matter -> numpy.ndarray
  • Eventually we'll want native matter objects in Python
• Bindings for msireg?
  • Need to coordinate with Sai
• Bindings for mzclustering?
  • Khoury Research Apprenticeship project

[kuwisdelu]

None of the above happened because US politics overwhelmed the author.

Let's try again.

Spring 2026

Cardinal 3.14 (Bioconductor 3.23 - expected April)

This update will be focused on usability and cleaning up the public API.

• Refactor core classes
  • Decouple MSImagingArrays and MSImagingExperiment
  • Use MSImagingArrays for processing spectra
  • Use MSImagingExperiment for downstream analysis
• Refactor spectral processing
  • Harmonize processing function implementations
  • Refactor process()
  • Refactor peakProcess()
  • Remove unnecessary generic functions
• Refactor results storage
• Refactor imzML import/export
• Refactor pixels() and features()

CardinalStats (under development)

Statistical learning methods will be moved to a new package.

• Prototype CardinalStats package
• Implement nearest neighbor methods (BiocNeighbors)
• Implement dimension reduction methods (BiocSingular)
• Implement clustering methods (bluster)