2024 Fulcher et al. tutorial
This tutorial analyzes paired single-cell RNA and proteomics data from C10 and SVEC cells in Fulcher et al. (2024). msmu processes Sage identifications and FlashLFQ peptide intensities and records the proteomics preprocessing steps. scanpy, PIMMS, muon, and MOFA provide RNA processing, protein imputation, and multi-omics analysis.
Workflow
Use docs/tutorials/2024_Fulcher as the working directory. For the main analysis, run 01 → 02 for proteins and 03 for RNA, then combine their outputs in 04. The protein and RNA notebooks download their inputs. Appendix A (workflow export and verification) and Appendix B (publication figures) are optional and can be skipped when following the main analysis.
| Notebook | Purpose | Main outputs |
|---|---|---|
| 01 Process proteomics data | Filter identifications, normalize peptides, and summarize proteins | MSV000089280.h5mu |
| 02 Impute protein data | Fill missing protein intensities with PIMMS | MSV000089280_imputed.h5mu |
| 03 Process scRNA-seq data | Map RNA counts to SDRF cell IDs, filter genes, and normalize | GSE201575.h5ad |
| 04 Integrate multi-omics data | Combine paired modalities, compute WNN UMAP, and fit MOFA | Fulcher_multimodal.h5mu, MOFA_model.hdf5 |
| Appendix A: Export and verify | Export the recorded workflow and compare protein-processing checkpoints | requirements.txt, script.py |
| Appendix B: Generate publication figures | Calculate differential expression and checkpoint comparisons, style figure panels, and export SVGs | Figures |
MSV000089280.h5mu is the processed checkpoint before imputation, with log₂-transformed, normalized, and summarized protein intensities. MSV000089280_imputed.h5mu also contains the PIMMS imputed layer used for WNN. Differential expression and MOFA use the observed protein matrix.
The verification section in Appendix A and the comparison figures in Appendix B require a separately prepared MSV000089280_debian.h5mu described in Appendix A. Appendix A compares the two checkpoints and fits both differential-expression analyses in the current Python session.
Environment setup
The recorded preprocessing environment uses Python 3.13.5 and msmu 0.4.3. Download requirements.txt for the exported package versions, including scanpy, PIMMS, muon, and MOFA dependencies.
References
Fulcher, J. M. et al. Parallel measurement of transcriptomes and proteomes from same single cells using nanodroplet splitting. Nature Communications 15, 10614 (2024). https://doi.org/10.1038/s41467-024-54099-z.