Morph2MolHISTOLOGY → SINGLE-CELL TRANSCRIPTOMEGitHub ↗

FROM MORPHOLOGY TO MOLECULES

Every cell has
a story.
Read it from H&E.

Morph2Mol predicts single-cell transcriptomes from routine histology, reading each cell’s morphology and spatial context across 18,580 genes.

University of Michigan · Weill Cornell Medicine

H&E of a pancreatic islet from donor HPAP-201Morph2Mol predicted INS and AMY2A expression for the same cellsH&EMorph2Mol
INS · isletAMY2A · acinarPancreas demo · HPAP-201 islet · workbench pipeline
25.4M

cells used for training

18,580

output genes

15

training tissue types

3.58M

held-out evaluation cells

01 / SELECTED APPLICATIONS

See single-cell transcriptomes
in spatial context.

From the architecture of an islet to the molecular programs of disease. Selected results from the Morph2Mol manuscript.

↗ EXPAND

02 / THE MODEL

One image.
Multiple layers of insight.

Morph2Mol combines morphology, spatial context and measured reference profiles. Its predictions are molecular scores—not measured molecule counts.

STAGE 01

Align

Cell-resolved histology representations align with scGPT transcriptomic features through contrastive learning.

→
STAGE 02

Contextualize

A spatial-attention decoder models cells and their neighbors to predict expression and generate cell embeddings.

→
STAGE 03

Retrieve & fuse

Gene-specific weights combine direct predictions with expression retrieved from reference-cell clusters.

03 / MORPH2MOL WORKBENCH

Your tissue.
A new perspective.

Upload a small H&E region, enter its physical resolution, and predict cell-level expression with cloud GPU inference.

Pancreas demo: H&E islet region

NO IMAGE? START HERE

Pancreas demo

A 521 × 521 px islet region from an HPAP donor with type 1 diabetes. Look for higher LTB in B cells next to CCL19-high endothelial cells at the islet edge. Upload it with the settings below, or click Pancreas demo inside the workbench to fill everything in.

  • 0.46 µm/pixel
  • Tissue: Pancreas
  • LTB · CCL19 · CCL21 · MX1 · CHGA · GCG
Download pancreas demo ↓
Checking inference service…PNG / JPEG / TIFF · up to 1024 × 1024 px · 20 MB

Free GPU time is limited per visitor. If the workbench reports that your ZeroGPU quota is used up, sign in to Hugging Face and open the workbench there ↗ to use your account’s quota.

↥

Connecting to the GPU workbench

Checking the status of the GPU workbench.

Open on Hugging Face ↗
01

Cell boundaries

Predicted nuclei and source-image coordinates.

02

Expression maps

View each gene relative to the other cells in your image. All tissues use within-image z-scores; download fused scores and z-scores as AnnData.

03

Cell embeddings

Stage 2 image features and spatial features as NPZ.

Research use only. Use de-identified images. Processing runs on Hugging Face; free GPU access is subject to availability and daily quotas. Whole-slide analysis is available through the command-line pipeline.

OPEN SCIENCE / MORPH2MOL

Morph2Mol: A Foundation Model for
Predicting Single-Cell Spatial Transcriptome
from Histology Images

Xin Luo*, Haoxuan Zeng* and colleagues

Research figure
Expanded research figure