Gradient Biotech

Welcome

The Pathology area of Gradient Biotech is a web workspace for computational pathology and whole-slide imaging research. Upload tissue slides or region images; generate tile manifests for browser viewing; run tissue detection, cell segmentation, and spatial quantification; inspect overlays in the slide viewer; and track reproducible pipeline runs.

What you can do today

CapabilityStatus
Slide uploadTIFF, OME-TIFF, PNG, JPEG, and scanner-format files
Tile generationThumbnail and multi-level tile manifest for web viewing
Tissue detectionTissue mask, region bounding boxes, focus QC
Cell segmentationNucleus/cell detection (threshold or watershed) with overlay and cell table
Object classifier trainingTrain a random-forest cell classifier from annotation-labeled regions; predict a class for every cell
Color deconvolutionH&E / H-DAB stain separation with per-stain intensity and DAB-positive area fraction
Spatial quantificationDensity, nearest-neighbor distances, infiltration phenotype, region metrics
Multiplex IF phenotypingPer-cell marker positivity, phenotype rules (e.g. CD8+ T cell, PD-L1+), co-expression and region composition
TMA dearrayingDetect tissue cores on a TMA slide and lay them out on a labeled grid
Spatial alignmentCoordinate-transform a compbio spatial transcriptomics run onto a pathology slide, with region-level expression summaries
Cohort comparisonGroup-level statistics (t-test/Mann-Whitney, ANOVA/Kruskal-Wallis, FDR) across sample groups
AI interpretationGrounded summaries citing computed metrics, with research disclaimers
OME-TIFF exportExport a processed slide as a pyramidal OME-TIFF for QuPath/Fiji/OMERO
Slide viewerPan/zoom tiles with tissue mask, cell, phenotype, density heatmap, and spatial-alignment overlay layers
Annotation importGeoJSON and QuPath-compatible region labels
Sample/cohort metadataPer-sample fields plus bulk CSV import
Run historyStudy-level pipeline run tracking with artifact links

How the product is organized

Every analysis lives inside a study — your pathology project container. From the pathology dashboard you create or open a study, then work through:

  1. Study home — upload slides and review run history
  2. Slide detail — generate tiles, run analysis pipelines, inspect the viewer
  3. Cohort — assign samples/groups, launch batch runs, compare metrics across slides
  4. Interpret — AI-assisted summaries of completed runs (open directly at /areas/pathology/studies/{id}/interpret)

Pipeline jobs run asynchronously. Launch jobs from the slide detail page and refresh when complete.

Who this is for

  • Translational and tumor biology researchers quantifying immune infiltration and tissue microenvironment metrics
  • Computational pathology groups running reproducible WSI pipelines at scale
  • Spatial biology labs grounding molecular assays in tissue morphology (spatial transcriptomics analysis lives in Computational Biology; pathology provides the tissue context layer)
  • Pharma tissue biomarker teams exploring histological endpoints with auditable run records

What this is not

  • Not a regulated clinical diagnostic or pathologist sign-out system
  • Not a passive slide viewer — overlays and metrics support quantitative research
  • Not automatic diagnosis — segmentation and quantification outputs are research metrics

Next steps

  • Use cases — scenario guides for slide quantification, TME infiltration, annotated compartments, and spatial omics context
  • Quick start — upload and analyze your first slide
  • Key concepts — studies, slides, runs, and artifacts
  • Study workflow — how the main sections fit together
  • AI interpretation — grounded summaries of completed runs