tmleAb

tmleAb: Targeted maximum likelihood estimation for antibody data

The tmleAb package estimates age-dependent antibody curves with ensemble machine learning, makes it easier to estimate marginally adjusted means from the curves, and compare marginally adjusted means between populations.

This package is mostly a convenience wrapper: the heavy lifting for machine learning is conducted using the SuperLearner package, and statistical estimation is conducted using the tmle package. In addition to these features, tmleAb also provides some methods for parametric models applied to antibody data (such as antibody acquisition models).

The package’s web interface is: http://ben-arnold.github.io/tmleAb/

An article that describes many of the methods implmented in the package is here: http://journals.plos.org/plosntds/article?id=10.1371/journal.pntd.0005616

With an associated GitHub repository here: https://github.com/ben-arnold/antibody-curves

Installation

Install tmleAb using the devtools package in R.

library(devtools)
install_github("ben-arnold/tmleAb")

The tmleAb package is in occasional development, so periodically check for updates. The package is not intended to be commercial-grade software, but it is intended to make it easier to use ensemble machine learning and targeted maximum likelihood estimation for the analysis of antibody measurements. If you use the package and run into issues and/or have comments, please write.

The package vignette (see Articles on the top of the package’s website) includes some details about the package and a few examples of how to use its main features.