EXPLANA: A user-friendly workflow for EXPLoratory ANAlysis and feature selection in cross-sectional and longitudinal...

Area of Science:

• Microbiome research
• Bioinformatics
• Statistical modeling

Background:

• Longitudinal microbiome studies (LMS) are growing but face analytical hurdles due to non-independent data and large datasets.
• Change analysis in LMS requires careful consideration of baseline as a reference, which varies between observational and interventional studies.

Purpose of the Study:

• To develop a robust feature selection workflow for cross-sectional and longitudinal microbiome data.
• To address the analytical challenges in LMS by integrating machine learning with change calculations.
• To provide a tool that simplifies the identification and explanation of variables related to outcomes.

Main Methods:

• Developed EXPLANA (EXPLoratory ANAlysis), a feature selection workflow for numerical and categorical data.
• Combined machine-learning algorithms with various change calculation methods and interpretation techniques.
• Generated an interactive report summarizing analytical methods and results both textually and graphically.

Main Results:

• EXPLANA demonstrated strong performance on simulated data, achieving an average area under the curve (AUC) of 0.91.
• The workflow significantly outperformed an existing analytical tool (AUC 0.95 vs. 0.56).
• Identified novel, order-dependent categorical feature changes, enhancing understanding of microbiome dynamics.

Conclusions:

• EXPLANA is a broadly applicable and effective tool for simplifying complex microbiome data analysis.
• The workflow aids in identifying statistically meaningful variables and explaining their relationship to outcomes in LMS.
• Facilitates deeper insights into microbiome dynamics and related factors through advanced feature selection.