Project Description

In this study, we have successfully replicated and built upon the analysis conducted in a highly influential paper in the field of breast cancer single-cell RNA sequencing (scRNA-seq), titled "Single-cell RNA-seq enables comprehensive tumour and immune cell profiling in primary breast cancer" that can be found with this link. Our research not only validates the feasibility of distinguishing between tumor and immune cells at the single-cell level but also sheds light on their characterization using targeted gene sets and unsupervised machine-learning techniques. These findings hold significant promise for advancing personalized medicine strategies in breast cancer. To enhance our analysis, we developed a preprocessing pipeline tailored to our specific research goals. By employing distinct and well-evaluated clustering approaches at each stage of cell separation, we aimed to optimize the accuracy and reliability of our results. Moreover, we incorporated advanced visualization techniques such as UMAP and t-SNE. These additions provided invaluable insights into the underlying organization and relationships within the data, enriching our overall analysis.

A comprehensive report of our findings and analysis can be found in the Project_Report.pdf file.

We present some example figures from the Project_Report.pdf below for easy access:

Visible clusters of carcinoma vs non-carcinoma cells utilizing T-SNE, UMAP, PCA on the Transcripts Per Million (TPM) counts matrix:

T-SNE, UMAP visualization of the Transcripts Per Million (TPM) counts matrix per cell type:

Results for gene expression analysis for ER+, HER2+, and TNBC marker genes for tumor cells and bulk tumors:

Data Availability

Both the publicly available datasets from the reference paper and our own data are available in this Google Drive Folder

Instructions

If you would like to run the code yourself, please follow the instructions below.

Library dependencies

• kneed
• matplotlib
• numpy
• pandas
• scanpy
• anndata
• seaborn
• sklearn
• optuna

Creating the workspace

1. Clone the repository
2. Download the reference paper data, as well as our data from the Google Drive Folder and place it in a folder named: datasets.
3. Run the notebooks in the following order 1. preprocessing.ipynb 2. dimensionalityred.ipynb 3. cellseperation.ipynb 4. R_genefu_results.ipynb 5. cancercellanalysis.ipynb 6. immunecellanalysis.ipynb

---