“Deconvolution Example”

This example loads blood DNA Methylation data from GEO, estimates the proportion of cell types, and compares them to known cell type proportions.

Load GEO blood methylation dataset measured on EPIC platform

from biolearn.data_library import DataLibrary

epic_facs_salas_18 = DataLibrary().get("GSE112618").load()

Estimate cell proportions using EPIC deconvolution model

from biolearn.model_gallery import ModelGallery
gallery = ModelGallery()

epic_facs_cell_prop = gallery.get("DeconvoluteBloodEPIC").predict(epic_facs_salas_18)

Estimating cell proportions using 'EPIC' platform reference...
Number of CpGs in reference: 600
Number of CpGs in methylation data intersecting with reference CpGs: 600
Running deconvolution...
Deconvolution complete!

Compare to known cell proportions measured with FACS (fluorescence-activated cell sorting)

import seaborn as sns
import matplotlib.pyplot as plt

sns.set_style("ticks")
plt.figure(figsize = (5, 5))

sns.scatterplot(x = epic_facs_salas_18.metadata.neutrophils_proportion.astype(float), y = epic_facs_cell_prop.loc["neutrophil"])
sns.scatterplot(x = epic_facs_salas_18.metadata.monocytes_proportion.astype(float), y = epic_facs_cell_prop.loc["monocyte"])
sns.scatterplot(x = epic_facs_salas_18.metadata.nk_proportion.astype(float), y = epic_facs_cell_prop.loc["nk_cell"])
sns.scatterplot(x = epic_facs_salas_18.metadata.bcell_proportion.astype(float), y = epic_facs_cell_prop.loc["b_cell"])
sns.scatterplot(x = epic_facs_salas_18.metadata.cd4t_proportion.astype(float), y = epic_facs_cell_prop.loc["cd4_t_cell"])
sns.scatterplot(x = epic_facs_salas_18.metadata.cd8t_proportion.astype(float), y = epic_facs_cell_prop.loc["cd8_t_cell"])

plt.legend(["Neutrophils", "Monocytes", "NK Cells", "B Cells", "CD4 T Cells", "CD8 T Cells"])
plt.xlabel("Known cell proportion")
plt.ylabel("Predicted cell proportion")
plt.title("Deconvolution Validation - FACS")

Text(0.5, 1.0, 'Deconvolution Validation - FACS')

Estimated memory usage: 118 MB