Joint principal trend analysis for longitudinal high-dimensional data.

Area of Science:

• * Computational biology and bioinformatics
• * Statistical modeling for high-dimensional data

Background:

• * Integrating diverse data sources is crucial for understanding complex biological processes.
• * Longitudinal, high-dimensional datasets present challenges in extracting meaningful biological insights.
• * Existing methods may not adequately capture shared latent trends across multiple datasets.

Purpose of the Study:

• * To develop a novel statistical method for analyzing multiple longitudinal, high-dimensional biological datasets.
• * To extract shared latent trends and identify key features indicative of dynamic biological processes.
• * To provide a robust tool for enhanced knowledge discovery in systems biology.

Main Methods:

• * Introduction of Joint Principal Trend Analysis (JPTA), a new statistical approach.
• * Application of JPTA to simulated datasets to validate its performance.
• * Utilizing JPTA on real-world biological data, including gene expression and transcriptional profiling.

Main Results:

• * JPTA successfully extracts shared latent trends from multiple high-dimensional datasets.
• * The method effectively identifies relevant features associated with biological processes.
• * Demonstrated utility in analyzing mammalian cell cycle gene expression and influenza infection data.

Conclusions:

• * Joint Principal Trend Analysis (JPTA) is an effective method for multi-dataset analysis in biology.
• * JPTA facilitates deeper understanding of dynamic biological processes through shared trend discovery.
• * The approach holds promise for advancing knowledge discovery in genomics and systems biology.