kornyok/pam50

A trained classifier to predict (breast) cancer patients PAM50 subtypes using gene expression data

PAM50

PAM50 package is a trained classifier that predicts PAM50 subtypes using patient's own gene expression data (NanoString, microarray, or RNA-seq). It can predict 5 subtypes for breast cancer patients: Luminal A, Luminal B, HER2, Basal, and Normal. Patients with Luminal A subtype usually have the best prognosis. Click here for more detailed description of each subtype.

gPAM50 uses 50 genes to make the prediction. As long as a patient has those 50 genes' expression data and Gene ID (Entrez ID), you can then predict PAM50 subtype for that patient. Yes, it works for a single patient's data! Interestingly, even the patient is not breast cancer patient, gPAM50 can still produce a prediction which may or may not be meaningful. But why don't you give it a try? Maybe you will discovery novel pathway or genes.

Project page: https://ccchang0111.github.io/PAM50/

Source code: https://github.com/ccchang0111/PAM50

Installation

You can install gPAM50 from github with:

# install.packages("devtools")
devtools::install_github("ccchang0111/gPAM50")

Tutorial

Magics can be found here: https://ccchang0111.github.io/PAM50/articles/PAM50_tutorial.html