📊 Available for scRNA-seq figure consulting — publication-ready figures with quantitative validation and reproducible code. Seurat or Scanpy.
→ biology2.net/services
End-to-end single-cell RNA-seq analyses on public datasets, prepared as a portfolio of reproducible workflows across cancer biology, immunology, development, and methodological showcases. Every project ships with: figures, a Jupyter or R notebook, and a short PDF report.
Tools: Scanpy (Python), Seurat (R), scVelo, Palantir, Harmony, BBKNN, Scanorama, scVI, CellChat, scib-metrics Reproducibility: conda environments, pinned versions, public data only Author: Marko Živanović — PhD, Principal Research Fellow, Institute for Information Technologies / University of Kragujevac
| # | Title | Tool | Dataset | Status |
|---|---|---|---|---|
| 01 | PBMC 3K — Reference scRNA-seq Pipeline | Scanpy | 10X PBMC 3K | ✅ |
| 02 | Breast Cancer Tumor Microenvironment | Scanpy | Wu et al. 2021, GSE176078 | ✅ |
| 03 | Glioblastoma Cellular States | Seurat | Neftel et al. 2019, GSE131928 | ✅ |
| 04 | Melanoma Immune Landscape | Seurat | Jerby-Arnon et al. 2018, GSE115978 | ✅ |
| 05 | Colorectal Cancer Single-Cell Atlas | Scanpy | Pelka et al. 2021 / Lee et al. 2020 | ✅ |
| 06 | RNA Velocity — Pancreatic Endocrinogenesis | scVelo | Bastidas-Ponce et al. 2019 | ✅ |
| 07 | Trajectory Inference — Hematopoiesis | Palantir + Scanpy | Setty et al. 2019 | ✅ |
| 08 | Multi-Sample Integration & Batch Correction | Scanpy + Harmony | Re-using #02 or #05 | ✅ |
| 09 | Cell-Cell Communication Analysis | CellChat (R) | Re-using #02 (TME) | ✅ |
| 10 | Cross-Platform Re-Analysis & Benchmarking | Scanpy + Seurat | TBD (published set) | ✅ |
Tick the box (☐ → ☑) when each project is fully shipped (figures + notebook + report).
upwork_portfolio/
├── README.md # this file
├── environment.yml # conda env (Python + R)
├── requirements.txt # pip-only alternative
├── Dockerfile # containerized alternative
├── .gitignore
├── _shared/
│ ├── scripts/ # helper scripts reused across projects
│ ├── styles/ # matplotlib/seaborn style for consistent figures
│ └── templates/ # notebook + report skeletons
├── docs/
│ └── portfolio_brief.md # short text descriptions ready for Upwork
└── NN_project_name/
├── README.md # project-specific brief
├── data/ # raw + processed data (gitignored, large)
├── notebooks/ # Jupyter / Rmd analysis notebooks
├── figures/ # PNG + TIFF outputs (300 dpi)
├── results/ # exported tables, h5ad/rds objects
└── report/ # PDF mini-report
# create the main environment
mamba env create -f environment.yml
mamba activate scportfolio
# verify
python -c "import scanpy; print(scanpy.__version__)"
Rscript -e 'library(Seurat); packageVersion("Seurat")'Each project must produce, before being marked complete:
- Hero figure — 1 visually striking publication-ready figure (UMAP, heatmap, velocity, etc.) saved as PNG (1200×900) for Upwork thumbnail + TIFF 300 dpi for manuscript-grade
- Supporting figures — 3–6 additional analytical figures
- Notebook — clean, narrative-driven, runnable end-to-end
- Mini-report — 2–3 page PDF (introduction → methods → results → biological interpretation)
- Project README — short summary + how-to-run + figure preview
- Public link — committed and pushed to GitHub
All datasets are public (GEO, Zenodo, Single Cell Portal, 10X Genomics) — proper citations are included in each project's README. Code in this repository is released under MIT.