# Code/data for reproducing the analyses reported in:
Madsen et al. 2021. Positive correlation between transcriptomic stemness and PI3K/AKT/mTOR signaling scores in breast cancer, and a counterintuitive relationship with PIK3CA genotype genotype. https://journals.plos.org/plosgenetics/article?id=10.1371/journal.pgen.1009876
## How-to-guide for breast cancer data
1. The analyses are split in two according to the two different breast cancer transcriptomic datasets used (METABRIC and TCGA). Further details for each are provided in the respective Wiki pages and in the following .Rmd files:
- METABRIC_breast_cancer_gene_signature_analysis_code_final.Rmd
- TCGA_breast_cancer_gene_signature_analysis_code_final.Rmd
2. To reproduce the workflow in its entirety, download the entire "data_package_PI3K_stemness_brca_paper.zip" file (go to "Files" to access) and make sure you set the resulting top folder as your working directory when you start R (recommended: use RStudio). You should then be able to simply open the .Rmd files and run them from the interface.
3. Note that there are three input directories. Briefly, the "input_Rueda_Caldas/" directory contains METABRIC-specific input data; the "input_files/" directory contains all the remaining input data required to reproduce the workflow; the "input_RDS/" directory contains .RDS file version of TCGA RNAseq dataset used for these analyses as well as the clinical data. These will be loaded automatically when running the TCGA notebook; if, however, the user prefers to download the most up-to-date versions of these datasets, there are instructions on how to do this within the notebook.
4. Exact package information can be found in the provided .html files, at the very beginning. Please refer to these files if the original code fails not run on your system as the .html files have all the output and code combined from the time the analyses were performed.
5. **POTENTIAL KNOWN ISSUE**: some users may get an error when they run the following line of code in the TCGA notebook
BRCARnaseqSE.1.all <- BRCARnaseq.all[, grepl("raw", colnames(BRCARnaseq.all))] %>%
mutate(Gene_id = rownames(BRCARnaseq.all))
To be able to continue, replace the above with the following:
BRCARnaseqSE.1.all <- as.data.frame(BRCARnaseq.all)[, grepl("raw", colnames(BRCARnaseq.all))] %>%
mutate(Gene_id = rownames(BRCARnaseq.all))
## Other data
Western blotting and RNA sequencing data from experiments with MCF10A overexpressing either empty vector (EV) or PIK3CA-H1047R are included in their respective folders, alongside the associated analysis scripts.
The raw sequencing data have been deposited on GEO under accession number GSE184277.
For inspection of the raw Western blotting images, please note that p110alpha and phospho-targets were blotted for separately (gels: 1, 3, and 5) to vinculin and the remaining total protein targets (gels: 2, 4 and 6). For protein size decoding, please refer to the PageRuler Plus pre-stained protein ladder (Fischer Scientific # PI26619) and the accompanying spreadsheet indicating the expected size for individual protein targets.
## Contact
If you encounter any other issues or have further questions regarding these analyses, please contact Dr Ralitsa Madsen (r.madsen@ucl.ac.uk).
