PAM50: PAM50 main function

In ccchang0111/PAM50: Breast Cancer's PAM50 prediction based on patient's RNA-seq data

Description

This is what you need for getting PAM50 prediction for each sample. It takes RNA-seq values (raw or CPM) and returns PAM50 subtype for each sample.

Usage

PAM50(df, cutoff = 0)

Arguments

df	dataframe, rows are genes columns are samples. Row names has to be EntrezID.
cutoff	numeric (0-1, default = 0), threshold value for calling the prediction. 0 means no cutoff, predicted subtype is whichever subtype gives the highest probability. 0.8 means the predicted subtype has to have probability > 0.8, otherwise 'NA' is the label.

Examples

library(Biobase)
library(tidyverse)
library(pamr)

df <- data.frame(assayData(es)$exprs)
df <- df[,1:10] ## select 10 samples
annotation <- fData(es)

rownames(df) <- annotation$EntrezID

df_pred <- PAM50(df)
knitr::kable(df_pred)

ccchang0111/PAM50 documentation built on May 31, 2019, 5:40 a.m.