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

Area of Science:

• Microbiome Research
• Bioinformatics
• Computational Biology

Background:

• Longitudinal microbiome studies (LMS) present analytical challenges, including non-independent data requiring mixed-effects models.
• The large volume of data necessitates exploratory analyses to identify factors related to outcomes.
• Change analysis in LMS is powerful but often lacks clear methodologies, especially distinguishing between observational and interventional study designs.

Purpose of the Study:

• To develop a robust feature selection workflow, EXPLANA (EXPLoratory ANAlysis), for longitudinal microbiome studies.
• To support both numerical and categorical data, and accommodate cross-sectional studies.
• To simplify the identification of statistically meaningful variables and their relationship to outcomes in LMS.

Main Methods:

• Combined machine learning methods with various change calculations and downstream interpretation techniques.
• Developed a feature selection workflow supporting numerical and categorical data.
• Integrated methods to generate an interactive report summarizing analyses and results.

Main Results:

• EXPLANA demonstrated strong performance on simulated longitudinal data with a balanced accuracy of 0.91.
• The workflow outperformed QIIME 2's feature-volatility tool (0.95 vs. 0.56 balanced accuracy).
• Identified novel order-dependent categorical feature changes, such as distinct effects for A_B versus B_A sequences.

Conclusions:

• EXPLANA is a broadly applicable tool that simplifies analytics for identifying features related to outcomes in LMS.
• The workflow effectively addresses challenges in analyzing longitudinal microbiome data.
• Provides a user-friendly, interactive report for clear interpretation of methods and results.