In tinahart97/ams597PROJ: This package implements different methods of pooling statistical evidence from multiple datasets.

knitr::opts_chunk$set(
  collapse = TRUE,
  comment = "#>"
)

library(MetaP)

Overall Projects

This function takes between 2 and 5 Data Frames and calculates p-values for each Biomarker to perform one of the 4 pooling tests below on.

The possible pooling method arguments are:

• "fisher" for Fisher

• "stouffer" for Stouffer

• "min" for Minimum P-value

• "max" for Maximum P-value

Usage: Project(x1, x2, x3, x4, x5, test)

Check input

This function checks if the inputs are vaild to use in our main function Checks that: All inputs are lists/ Data Frames Same number of columns in each Data Frame Rows 2 to p are biomarkers Checks that the group membership column will take at least two unique values

Usage: Check.input(frames)

GroupDifference

This function takes a Data Frame that can be normally distributed or not normally distributed. For 2 groups we will perform a two sample t-test or Wilcoxon rank sum test based on Normality. For more than 2 groups we will perform ANOVA or Kruskal Wallis test based on Normality.

Usage: GroupDifference(x)

The return is an output of pvalues for each biomarker

Pool.Fisher

This function sums the log-transformed p-values, following a chi-squared distributions with 2k degrees of freedom

Usage: pool.fisher(x)

The return is a vector of p-values, one for each biomarker

Pool.Stouffer

This function sums the inverse normal p-values, following a standard normal distribution

Usage: pool.stouffer(x)

The return is a vector of p-values, one for each biomarker

This function follows a beta distribution with degrees of freedom α=1 and β=k

Usage: pool.min(x)

The return is a vector of p-values, one for each biomarker

Pool.max

This function follows a beta distribution with degrees of freedom α=k and β=1

Usage: pool.max(x)

The return is a vector of p-values, one for each biomarker

tinahart97/ams597PROJ documentation built on May 8, 2019, 12:55 a.m.