SPatial Inference of Cellular Environments via Metabolic Modeling
Work in progress — manuscript in preparation.
SPICEM is a spatial metabolic modeling pipeline for tissue transcriptomics. Given spatially resolved single-cell/spot data, it builds local multi-cell community metabolic models and analyzes the resulting flux solutions along two complementary axes:
- Exchange fluxes — metabolite secretion/uptake between spatially proximal cells, used to map cell-cell metabolic interactions and how they shift across clinical groups.
- Intracellular fluxes — each cell's own internal metabolic activity (individual reactions and, via reference-model subsystem annotations, whole pathways), used to characterize cell-type-specific metabolic phenotypes and their spatial organization.
This repository accompanies ongoing work applying the pipeline to human kidney tissue, relating spatial metabolic dysfunction to disease-relevant cell types. The associated manuscript is currently in preparation and unpublished; this repo is shared to document active development.
Spatial metabolic modelling maps GWAS risk to cell-type-specific dysfunction in diabetic kidney disease
The pipeline runs as a set of Python modules driven from a Jupyter notebook
(spicem_human_cohort.ipynb), with MATLAB/COBRA scripts (model_building/)
handling community model construction and flux extraction upstream.
Note:
model_building/contains placeholder files in this shared copy of the repository — filenames and each script's documented purpose are preserved so the pipeline's structure is clear, but the MATLAB implementations are withheld while the associated manuscript is in preparation. Contact the author for access.
from pipeline import AnalysisConfig, run_analysis_pipeline
cfg = AnalysisConfig(base_dir="<path-to-region-data>", conditions=["Sample_Condition"])
results = run_analysis_pipeline(cfg)See the notebook for the full cohort workflow (per-patient runs, cohort aggregation, statistical comparisons, and figure generation).
Pre-publication, active development. All rights reserved — no license is granted for reuse or redistribution at this time. This will be revisited upon publication.
Lokanand Koduru, Senior Scientist at the Genome Institute of Singapore (GIS), Agency for Science, Technology and Research (A*STAR).