Developing an Embedding, Koopman and Autoencoder Technologies-Based Multi-Omics Time Series Predictive Model (EKATP)...

Area of Science:

• Systems Biology
• Bioinformatics
• Computational Biology

Background:

• Predicting multi-omics time series is crucial for disease occurrence and health detection.
• High-dimensional, nonlinear, and noisy characteristics of multi-omics data pose significant prediction challenges.

Purpose of the Study:

• To propose an innovative multi-omics time series predictive model.
• To address the difficulties in predicting complex biological time series data.

Main Methods:

• Developed an Embedding, Koopman, and Autoencoder technologies-based multi-omics time series predictive model (EKATP).
• Evaluated EKATP using genomics (chaotic), proteomics (oscillating), and metabolomics (flow) time series data.

Main Results:

• EKATP demonstrated substantial improvements in prediction accuracy.
• The model showed enhanced robustness and generalizability for multi-omics time series.
• Successfully predicted future states across diverse biological data types.

Conclusions:

• EKATP offers a powerful new tool for systems biologists.
• The model effectively handles the complexities of multi-omics time series.
• EKATP advances the prediction of disease occurrence and health status from biological data.