README.md

In lc5415/HDATDS: Useful function for TDS project

HDATDS

Package with useful functions for the MSc HDA&ML Translational Data Sciences and Computational Epidemiology projects

Functions built by March 31st:

• BHSCalculator()
• formatBio()
• quantile_check() --> mostly internal use
• Analysis(): performs cross-validation for any given model and given dataset or combination of datasets

Install in your machine (from RStudio): Note: the imperial HPC RStudio session cannot install xgboost which is a dependency of this package, so do not try to install this package from the HPC RStudio.

library(devtools)
library(dplyr)
devtools::install_github("lc5415/HDATDS")
library(HDATDS)

Install in your machine (Terminal): I would recommend installing this package in a new environment given dependecy issues may arise otherwise.

conda create --name NewProjectEnv
conda activate NewProjectEnv
conda install -c lucha6 r-hdatds

Example code:

# Original biomarker dataframe
data("bio.example")
data("cov")
# Load biomarkers and covariates data frames (bio and cov) and merge them by ID
bio = merge(bio.example, cov[,c("ID","age_cl","gender")], by = "ID")
ids = bio$ID # keeping explicit copy of IDs
rownames(bio) = bio[,1] # assuming ID is the first column
bio = bio[,-1]
bio$age_cl = as.factor(bio$age_cl)
bio$gender = as.factor(bio$gender)

bio = bio[complete.cases(bio), ] #function does not handle NAs internally

# Run BHS calculation using paper reference
scores_paper = BHSCalculator(bio, "Paper", stratified = TRUE, bySystems = TRUE)

lc5415/HDATDS documentation built on April 27, 2020, 6:04 a.m.