Skip to content

VIPP 0.16.0a1

VIPP 0.16.0a1 brings measurement collection, descriptive statistics and editable plots into one connected Results Workspace. It also adds explicit 2D CellProfiler compartment tools and clearer workflow navigation.

Use the canonical GitHub release for the wheel, source archive, Windows installer, separate Apple Silicon/Intel macOS packages and matching checksums, or the exact PyPI version. Desktop installers are unsigned alpha builds; macOS packages are also unnotarized and CPU-only. Follow the installation instructions.

Explore measurements without leaving the workflow

Results Workspace places Data, Summary and Plots beside their settings. The persistent selection bar identifies the workflow, exact table output, Statistics node and plot you are editing. Browse any open workflow, maximize the window, and use Show node to return to the selected graph node.

Opening the workspace does not create an analysis. Adding a summary or plot creates an ordinary saved node; changes stay synchronized with its inspector and other windows. Browsing, searching rows and hiding columns do not filter the scientific data or change full-table exports.

  • Statistics reports counts, mean, median, sample SD, quartiles, IQR, minimum, maximum and sum. Choose objects, image averages or sample averages, with explicit within-sample weighting and inclusion counts.
  • Plot Results offers grouped points, histograms, cumulative distributions and scatter plots. Choose original measurements or a summary table explicitly; a summary row does not become an original object again.
  • Set automatic or custom numeric tick spacing, retain whole-number count axes, and export sized PNG/TIFF or SVG/PDF figures. Busy indicators keep their space, setup messages explain missing choices, and out-of-date results cannot be exported as current.
  • Use Increase Decimal and Decrease Decimal to change table display precision from the initial 3 decimal places. Hover for the original value; calculations, sorting, searching and exports retain full-precision data.

These are descriptive tools. They do not add significance tests, ANOVA, p-values or confidence intervals. Counts of cells or images are not proof of independent biological replication. See the Statistics methods and counts.

Collect results across images

Collect batch measurements reviews saved table outputs and their recorded identity before combining them. Add condition/sample annotations, review exclusions, then export CSV, TSV or an Excel workbook directly. You do not need to rerun the images or open another workflow first.

The image inventory keeps missing, failed, excluded and valid empty results distinct. Excel includes measurements, an image summary and collection information; CSV/TSV offers a separate image-summary file. Save the optional typed .vipp-results.json dataset to reopen through Table Source while preserving VIPP's types, units and collection records.

Grow and measure seeded compartments

Grow Regions from Seeds — CellProfiler Propagation uses Centrosome 1.3.4 on CPU to grow labels from a guidance image, seed labels and a Boolean mask. It requires one aligned YX plane and does not silently normalize the guidance image. See Grow regions from seeds.

Six additional nodes expose the recorded CellProfiler 4.2.6 compartment profile: Gaussian smoothing, minimum-cross-entropy thresholding, shape-based nuclei, propagation-seed preparation, cell-region finishing and cytoplasm extraction. The worked paper workflow explains the explicit scaling, connections and compartment measurements.

This is a specific 2D profile, not every CellProfiler setting or a new 3D Propagation method. Existing Marker-Controlled Watershed remains the 3D route. Agreement with the reference implementation, reproduction of published measurements and biological validity remain separate claims; the validation status records the evidence and unresolved differences.

Easier everyday use

  • Name nodes in the inspector or with Rename…. Automatic table/summary/plot names describe their data or recipe. Names are saved and undoable; renaming does not change calculations or figure titles.
  • Use the three-dot Workflow actions menu beside Save for examples, Save workflow as… and exports, separate from settings.
  • Windows setup shows a live installation log. Scroll up to review earlier output; Jump to latest resumes following.
  • Dialog footer actions follow the platform's order, with Cancel/Close rightmost on Windows and before the main action on macOS.

What to try first

  1. Open Morphology & Intensity Plots from the example chooser. Its 60 synthetic objects demonstrate morphology, intensity, histograms and scatter.
  2. Open Results Workspace from a measurement node. Add a summary, choose its observation level, then deliberately plot either its summary rows or the original measurements.
  3. For seeded segmentation, follow the focused guide above. The Exhaustive Inspector Showcase also contains synthetic Propagation and compartment lanes for reviewing controls; these are not paper-data validation.

Compatibility and scope

Workflow, batch-config and manifest schemas remain 6. New operations, Statistics recipes and native measurement datasets still require a compatible runtime; equal schema numbers do not make them usable in older VIPP releases.

Older Summarize Measurements nodes keep their version-1 calculations until an explicit upgrade. New Statistics nodes use version 2, including undefined sample SD when fewer than two eligible observations remain. Review counts, units, weighting and exclusions when upgrading; do not assume an old and a new recipe have identical meanings.

Preserve original workflows, inputs, environments and decisive outputs. Recalculate representative data and regenerate Python exports for the installed version. The new descriptive and compartment operations are CPU methods; this release scope does not broaden GPU qualification.

The qualification declaration separates changed-domain checks from carried-forward evidence. Reference comparisons and synthetic examples do not establish biological validity.

Registration, image-comparison metrics and template matching remain later 0.16-series work. They are not included in this first alpha's feature set.

Full source comparison: 0.15.0a5 → 0.16.0a1