GRN Parameter Inference

Target State Setup

Define the desired gene expression levels

Quick Presets:

▼ Progenitor TFs

▼ Lineage TFs

▶ Ligands

▶ Receptors

Gene Weights

Select which genes must match the target

Only selected genes contribute to fitness. Unselected genes can be anything.

Progenitor TFs

Lineage TFs

Ligands

Receptors

Selected: 0 genes

Optimization Config

Global parameters

Gene knockouts

Morphogen per-lineage

Uniform application

Apply morphogens to all cells equally (no spatial gradient)

Gene modifiers (drugs/inhibitors)

Optimize continuous modifiers (0=inhibited, 1=normal, 2=overexpressed)

Minimize interventions

Find smallest set of modifiers needed (L1 regularization)

Sparsity Penalty: 0.10

Higher = fewer interventions (but may sacrifice fitness)

Require homogeneous population

ALL cells must match target (optimizes worst cell, slower but ensures uniformity)

Max Generations: 200

Population Size: 20

Simulation Time: 12h

Target Lineage:

Select a lineage to optimize for cells in that spatial region

Optimization Progress

Ready

Generation 0 / 200

Avg Fitness -

Worst Cell -

Correlation -

Target vs Simulated

Prog TF Lin TF Ligand Receptor

Top Errors

Genes with largest mismatch

Expression by Gene Group

Target Simulated

Found Parameters

Run optimization to see results

Knockouts

Morphogen Strengths

Gene Modifiers

0=inhibited, 1=normal, 2=overexpressed

Sensitivity Analysis

Identify which parameters have biggest impact

Run optimization first

Stability Check

Verify if the found state is a stable attractor

Run optimization first

Threshold: 0.10

Optimization Progress

Target vs Simulated

Top Errors

Expression by Gene Group

Edit Timepoint: t = 0h

Import Cell State