README.md

In hk1785/mMeta: Multi-marker meta analysis (mMeta) and Adaptive multi-marker meta-analysis (aMeta)

R package: mMeta

Title: Multi-marker meta analysis (mMeta) & Adaptive multi-marker meta-analysis (aMeta)

Version: 1.0

Date: 2021-2-11

Author: Hyunwook Koh

Maintainer: Hyunwook Koh [e-mail]

Description: This R package provides facilities for multi-marker meta analysis (mMeta) and adaptive multi-marker meta-analysis (aMeta) which conduct meta-analyses to combine multiple studies throughout multiple related markers.

NeedsCompilation: No

Depends: R(>= 4.0.3)

Imports: forestplot, MASS, vegan

License: GPL-2

URL: https://github.com/hk1785/mMeta

Reference

• Koh H, Zhao N. (2021) Meta-analysis methods for multiple related markers: applications to microbiome studies with the results on multiple α-diversity indices. Statistics in Medicine 40(12), 2859-2879.

Troubleshooting Tips

If you have any problems for using this R package, please report in Issues (https://github.com/hk1785/mMeta/issues) or email Hyunwook Koh ([e-mail]).

Prerequites

devtools

install.packages("devtools")

forestplot

install.packages("forestplot")

MASS

install.packages("MASS")

vegan

install.packages("vegan")

Installation

library(devtools)
install_github("hk1785/mMeta", force=T)

Manual

This R package includes two core functions, mMeta.aMeta and mMeta.aMeta.plot, and one summary data set, HIV.Sum.Data. Please find the details below.

:mag: mMeta.aMeta

Description

This function conducts the meta-analysis using mMeta and aMeta to combine multiple studies throughout multiple related markers.

Usage

mMeta.aMeta(est, std.err, tau0=FALSE, n.perm=5000, seed=NULL)

Arguments

• est - A data.frame (or matrix) of study-specific estimates (e.g., regression coefficients). Rows are studies (row names are study names) and columns are markers (column names are marker names). 'est' can contain missing elements (NAs), but the locations of NAs should be matched between 'est' and 'std.err'. The number of studies (i.e., the number of rows) should be greater than or equal to 10.
• std.err - A data.frame (or matrix) of study-specific standard errors. Rows are studies (row names are study names) and columns are markers (column names are marker names). 'std.err' can contain missing elements (NAs), but the locations of NAs should be matched between 'est' and 'std.err'. The number of studies (i.e., the number of rows) should be greater than or equal to 10.
• tau0 - An indicator to set up the heterogeneous variance exists under the null (TRUE), which is the traditional random effects meta-analysis, or the heterogeneous variance does not exist under the null (FALSE), which is the Han and Eskin’s modified random effects meta-analysis. Default is TRUE.
• n.perm - A number of permutations. Default is n.perm=5000. If the number of studies < 15, all possible permutations are implemented and 'n.perm' is ignored.
• seed - A seed number to obtain the same outcomes repeatedly. Default is NULL for no seed number. If the number of studies < 15, 'seed' is ignored as all possible permutations are implemented.

Values

Pooled estimates, 95% confidence intervals, p-values for each marker and across markers using mMeta, and a p-value across markers using aMeta

References

• Koh H, Zhao N. (2021) Meta-analysis methods for multiple related markers: applications to microbiome studies with the results on multiple α-diversity indices. Statistics in Medicine 40(12), 2859-2879.

Example

Import requisite R packages

library(forestplot)
library(MASS)
library(vegan)
library(mMeta)

Import example summary data

data(Sum.Data)
est <- Sum.Data$est
std.err <- Sum.Data$std.err

Fit mMeta and aMeta

out1 <- mMeta.aMeta(est, std.err, seed=123)
out1

Import summary data for 15 HIV studies and 6 alpha-diversity indices

data(HIV.Sum.Data)
est <- HIV.Sum.Data$est
std.err <- HIV.Sum.Data$std.err

Fit mMeta and aMeta

out2 <- mMeta.aMeta(est, std.err, seed=123)
out2

:mag: mMeta.aMeta.plot

Description

This function draws a forest plot for each marker, and across markers using mMeta and aMeta.

Usage

mMeta.aMeta.plot(mMeta.aMeta.out)

Arguments

• mMeta.aMeta.out - An output obtained using the mMeta.aMeta function.

Values

A forest plot for each marker, and across markers using mMeta and aMeta

References

• Koh H, Zhao N. (2021) Meta-analysis methods for multiple related markers: applications to microbiome studies with the results on multiple α-diversity indices. Statistics in Medicine 40(12), 2859-2879.

Example

Import requisite R packages

library(forestplot)
library(MASS)
library(vegan)
library(mMeta)

Import example summary data

data(Sum.Data)
est <- Sum.Data$est
std.err <- Sum.Data$std.err

Fit mMeta and aMeta

out1 <- mMeta.aMeta(est, std.err, seed=123)
out1

Draw a forest plot

mMeta.aMeta.plot(mMeta.aMeta.out=out1)

Import summary data for 15 HIV studies and 6 alpha-diversity indices

data(HIV.Sum.Data)
est <- HIV.Sum.Data$est
std.err <- HIV.Sum.Data$std.err

Fit mMeta and aMeta

out2 <- mMeta.aMeta(est, std.err, seed=123)
out2

Draw a forest plot

mMeta.aMeta.plot(mMeta.aMeta.out=out2)

:mag: HIV.Sum.Data

Description

A list of regression coefficients and their standard errors for the 15 human microbiome studies for the association between HIV status and each of the 6 alpha-diversity indices, Richness (Observed), Shannon, Simpson, PD, PE and PQE.

References

• Koh H, Zhao N. (2021) Meta-analysis methods for multiple related markers: applications to microbiome studies with the results on multiple α-diversity indices. Statistics in Medicine 40(12), 2859-2879.

Example

Import requisite R packages

library(mMeta)

Import summary data for 15 HIV studies and 6 alpha-diversity indices

data(HIV.Sum.Data)
est <- HIV.Sum.Data$est # Regression coefficients
std.err <- HIV.Sum.Data$std.err # Standard errors

hk1785/mMeta documentation built on Dec. 20, 2024, 3:27 a.m.