ComBatFamQC
The ComBatFamQC package is an innovative solution designed to enhance the process of batch effect diagnostics, harmonization, and post-harmonization downstream analysis.
By providing interactive qualitative visualization alongside rigorous statistical testing, this package empowers users to effectively identify potential batch effects in their data. Additionally, ComBatFamQC includes a suite of built-in harmonization techniques that streamline the harmonization process, allowing researchers to achieve optimal results without requiring extensive technical expertise. The four harmonization techniques supported are: 1) Original ComBat, 2) Longitudinal ComBat, 3) ComBat-GAM, and 4) CovBat. Furthermore, ComBatFamQC allows users to create interactive life-span age trend plots of individual brain features with estimated age-adjusted centiles. The package also provides functions to generate covariate-corrected residuals, which can be used as inputs for downstream machine learning models.
The package leverages the capabilities of R Shiny to offer interactive visualizations for both batch effect diagnostics and age trend analysis, presenting data in an accessible format that enhances interpretability.
To cater to a diverse user base, ComBatFamQC includes two user-friendly command-line interfaces (located in the inst/ directory) that guide users through different stages of the processing pipeline:
The CombatQC_CLI.R script focuses on the batch effect diagnostics and harmonization stage, providing interactive tools to diagnose and adjust for batch effects effectively.
The post_CLI.R script is dedicated to the post-harmonization stage, where users can visualize age trends across various brain structures and generate a residual dataset that excludes the effects of unwanted covariates.
Together, these components make ComBatFamQC a comprehensive and accessible tool for researchers working with complex datasets in neuroimaging and other related fields.