Piecewise Multivariate Modelling of Sequential Metabolic Profiling Data
Author Information
Author(s): Mattias Rantalainen, Olivier Cloarec, Timothy Ebbels, Torbjörn Lundstedt, Jeremy K Nicholson, Elaine Holmes, Johan Trygg
Primary Institution: Imperial College London
Hypothesis
Can a piecewise multivariate modelling approach effectively model and predict time-related changes in metabolic profiling data?
Conclusion
The proposed method is effective for modelling and prediction of short and multivariate time series data.
Supporting Evidence
- The method allows for easy interpretation of time-related variation in the data.
- It provides a competitive complement to commonly applied multivariate methods such as OPLS and PCA.
- The approach can be used for monitoring disease stage over time.
Takeaway
This study introduces a new way to look at how biological systems change over time by using a special method to analyze data from experiments.
Methodology
The study used a supervised multivariate modelling approach with piecewise Orthogonal Projections to Latent Structures (OPLS) models to analyze time-related changes in metabolic data.
Limitations
The method does not exploit autocorrelation structures in the time-series and may struggle with unsynchronized samples.
Participant Demographics
The study involved ten rats in a renal toxicity study using mercury II chloride.
Digital Object Identifier (DOI)
Want to read the original?
Access the complete publication on the publisher's website