2025: Out now at Cancer Letters!
We developed a novel computational framework called DECODEM (DEcoupling Cell-type-specific Outcomes using DEconvolution and Machine learning) that can systematically assess the roles of the diverse cell types in the tumor microenvironment (TME) in a given phenotype from bulk transcriptomics. In this work, we investigate the association of the cell types in breast cancer TME (BC-TME) to patient response to neoadjuvant chemotherapy (responder vs. non-responder). The framework is divided into two steps:
- Deconvolution: we use CODEFACS to deconvolve the bulk gene expression into nine cell-type-specific gene expression profiles encompassing malignant, immune, and stromal cell types.
- Machine Learning: we use a machine learning (ML) pipeline to build nine cell-type-specific predictors of chemotherapy response using the deconvolved expression profiles.
The output of the framework is the likelihood scores that the patients will respond to chemotherapy. We then rank the cell types within the BC-TME based on their predictive power (in terms of AUC, AP and DOR), identifying "prominent" cell types that provide improvements over the bulk mixture. We further validate the prominent cell types in multiple independent BC cohorts encompassing both bulk and single-cell (SC) transcriptomics.
AUC = Area under the receiver operating characteristics curve,
AP = Average precision, equivalent to the area under the precision-recall curve,
DOR = Diagnostic odds ratio
Figure: The full analysis pipeline for DECODEM and DECODEMi
Furthermore, we investigate the interactions between different cell types in two ways:
- Multi-cell-ensemble: we incorporate the expression profiles of the top predictive cell types to boost the predictive power even further, yielding the best performance for an ensemble of immune and stromal cell types across two independent cohorts.
- DECODEMi: we extended DECODEM to DECODEMi ('i' stands for interaction) where we use the inferred cell-cell interactions (CCIs) (by using LIRICS) to identify the cellular communications that influence chemotherapy response in BC.
Our findings in breast cancer highlight the considerable predictive powers of the immune and stromal cells in the TME as well as denote key CCIs that are strongly predictive of chemotherapy response.
The deconvolution (and CCI inference) stage was performed on NIH Biowulf environment using R and Rslurm.
The ML predictors were developed on MacOS using python and further tested on linux (on HPC). The ML scripts can be run interactively using a python IDE or on command line as python script_name.py. Complementary analyses i.e., data preprocessing, enrichment analysis, CCI validation in SC and some plot generation were performed locally using R on RStudio.
Dependencies for python scripts:
python >= 3.10
numpy >= 1.23
pandas >= 1.4
scikit-learn >= 1.1
xgboost >= 1.6.1
pickle >= 3.0
matplotlib >= 3.7
seaborn >= 0.12
tqdm >= 4.63
lifelines >= 0.27
pickle == 4.0 Dependencies for R scripts:
R >= 3.6
tidyverse >= 1.3
plyr >= 1.8
rtracklayer >= 1.57
GenomicFeatures >= 1.50
clusterProfiler >= 4.6
biomaRt >= 2.54
msigdbr >= 7.5
GSVA >= 2.4
PRROC >= 1.3
rstatix >= 0.7
ggpubr >= 0.6
seurat >= 5.1.0
glue >= 1.6
Matrix >= 1.6
CellChat >= 2.1 All the results presented in the above manuscript can be reproduced by using the scripts provided in analysis. The assumption is that the different bulk expression datasets have already been deconvolved and put in the designated directories within data.
Running deconvolution with CODEFACS and LIRICS
The scripts for CODEFACS and LIRICS should respectively be put in analysis/deconvolution/CODEFACS and analysis/deconvolution/LIRICS. The cell type signature should be in data/celltype_signature.
- Deconvolution using CODEFACS was run by using the
slurmscripts in analysis/deconvolution/job_scripts. - The
slurmscripts were run on the NIH HPC system, Biowulf. - CCI inference using LIRICS was run by using the scripts in analysis/deconvolution/LIRICS.
All datasets should be deposited in data using the structure outlined. To process the deconvolved data into the desired formats, use the scripts in analysis/preprocessing.
Examples of some processed datasets are provided in data/TransNEO and data/BrighTNess.
model_transneo_cv_vX.py: performs the cross-validation analysis using the TransNEO cohort.predict_sammut_validation_vX.py: trains the cell-type-specific / multi-cell-ensemble predictors using TransNEO and validates on the ARTemis + PBCP cohort.predict_brightness_validation_vX.py: trains the cell-type-specific / multi-cell-ensemble predictors using TransNEO and validates on the BrighTNess cohort containing triple negative breast cancer (TNBC) patients.predict_zhang_sc_validation_vX.py: trains the cell-type-specific predictors using TransNEO and validates on the Zhang et al. SC cohort of TNBC patients.predict_bassez_sc_validation_vX.py: trains the cell-type-specific predictors using TransNEO and validates on the Bassez et al. SC cohort of TNBC patients.stratify_tcga_validation_vX.py: trains the cell-type-specific predictors using TransNEO and stratifies survival on the TCGA-BRCA cohort.- files with
_looin their name : performs hyperparameter tuning using a leave-one-out cross-validation.
If svdat = True in the scripts, the predictions will be saved in data/TransNEO/transneo_analysis/mdl_data (in .pkl format).
model_transneo_lirics_cv_vX.py: performs the cross-validation analysis using TransNEO and extracts the corresponding top predictive CCIs.predict_sammut_lirics_validation_vX.py: trains the CCI-based predictor using TransNEO, validates on ARTemis + PBCP and extracts the corresponding top predictive CCIs.predict_brightness_lirics_validation_vX.py: trains the CCI-based predictor using TransNEO, validates on BrighTNess and extracts the corresponding top predictive CCIs.predict_zhang_lirics_sc_validation_cellchat_vX.R: validates the top predictive CCIs in TNBC extracted by DECODEMi with Zhang et al. SC cohort (CCIs extracted by CellChat v2) and generates Figs. 4G-H.
If svdat = True in the scripts, the predictions will be saved in data/TransNEO/transneo_analysis/mdl_data (in .pkl format).
The enrichment analyses results and the figures (or panels) in the manuscript can be reproduced using the scripts in analysis/enrichment_and_figures.
run_enrichment_top_cell_types_vX.R: performs cell-type-specific GSEA analysis and generates Fig. 3G.enrichment_cd4_cd8_tcells_vX.R: performs GSVA analysis for CD4+/CD8+ T-cells, estimates their predictive power and generates Supp. Figs. 6G-J.get_abundance_response_corr_vX.py: performs an association analysis between cell type abundance and chemotherapy response, and generates Supp. Fig. 5.
If svdat = True in the scripts, the figure panels will be saved in data/plots (in .pdf format, DPI = 600).
Fig. 1D was generated using Biorender (Dhruba, S. R. (2025)). The remaining figures were generated using the following scripts in analysis/enrichment_and_figures (and further polished using Adobe Illustrator):
generate_plots_ctp_vX.py: generates Figs. 1A-B, 2, 3A-F, Supp. Fig. 2-3.generate_plots_cci_vX.py: generates Figs. 4A-F, Supp. Figs. 8A-D.generate_plots_sc_surv_vX.py: generates Figs. 1C, 5, Supp. Figs. 10-11.explore_drug_by_icd_vX.py: generates Supp. Fig. 7.make_benchmark_figures_vX.R: generates Supp. Fig. 1.
if svdat = True in the scripts, the figures will be saved in data/plots (in .pdf format, DPI = 600).
The final figures are provided in figures.
If you use DECODEM or DECODEMi in your research / application, please cite the following:
Dhruba, S. R. et al. (2025), Enhanced prediction of breast cancer patient response to chemotherapy by integrating deconvolved expression patterns of immune, stromal and tumor cells, Cancer Letters. DOI: https://doi.org/10.1016/j.canlet.2025.218101
Saugato Rahman Dhruba (dhruba018@gmail.com)
Cancer Data Science Lab, NCI, NIH