Skip to main content
🔍

Deconvolution of PBMC Data

Introduction​

This demo shows the Statescope workflow on PBMC bulk data. The process includes:

  1. Importing libraries and loading a bulk dataset.
  2. Initializing the Statescope model.
  3. Running deconvolution, refinement, and state discovery.
  4. Visualizing the results with:
    • A heatmap of cell type fractions,
    • A plot of cophenetic coefficients,
    • A bar plot of state loadings.

Step 1: Import Libraries​

import Statescope.Statescope as scope
import pandas as pd

Step 2: Load the Bulk Dataset​

# Read bulk dataset GSE263756
Bulk = pd.read_csv(
    'https://github.com/tgac-vumc/StatescopeData/raw/refs/heads/main/Bulk/PBMC/GSE263756.txt', 
    sep='\t', 
    index_col='Geneid'
)
Bulk.head(3)

Step 3: Initialize the Statescope Model​

Statescope_model = scope.Initialize_Statescope(Bulk, TumorType='PBMC', Ncelltypes=7)

Step 4: Run Deconvolution​

Statescope_model.Deconvolution()

Console output shows that deconvolution completed successfully.

Step 5: Run Refinement​

Statescope_model.Refinement()

Console output shows that refinement completed successfully.

Step 6: Run StateDiscovery​

Statescope_model.StateDiscovery()

Console output shows that state discovery completed successfully.

Step 7: Visualize Results​

7.1: Heatmap of Cell Type Fractions​

from Statescope.Statescope import Heatmap_Fractions
Heatmap_Fractions(Statescope_model)

Heatmap Fractions

7.2: Cophenetic Coefficients Plot​

from Statescope.Statescope import Plot_CopheneticCoefficients
Plot_CopheneticCoefficients(Statescope_model)

Cophenetic Coefficients

7.3: Bar Plot of State Loadings​

from Statescope.Statescope import BarPlot_StateLoadings
BarPlot_StateLoadings(Statescope_model)

Console output: "StateLoadings matrix extracted successfully. Shape: (18638, 27)"

Bar Plot State Loadings