Hierarchical Cluster-based Partial Least Squares Regression (HC-PLSR) is an efficient tool for metamodelling of nonlinear dynamic models
2011

A New Method for Modeling Complex Biological Systems

Sample size: 59049 publication Evidence: high

Author Information

Author(s): Tøndel Kristin, Indahl Ulf G, Gjuvsland Arne B, Vik Jon Olav, Hunter Peter, Omholt Stig W, Martens Harald

Primary Institution: Norwegian University of Life Sciences

Hypothesis

A more accurate mapping can be obtained by locally linear or locally polynomial regression.

Conclusion

HC-PLSR is a promising approach for metamodelling in systems biology, especially for highly nonlinear or non-monotone parameter to phenotype maps.

Supporting Evidence

  • HC-PLSR outperformed both polynomial PLSR and OLS regression in all test cases.
  • The advantage of HC-PLSR was largest in systems with highly nonlinear functional relationships.
  • HC-PLSR can flexibly adjust to suit the complexity of dynamic model behavior.

Takeaway

This study shows a new way to model complex biological systems that helps predict how changes in parameters affect outcomes, especially when those relationships are complicated.

Methodology

The study used a new method called HC-PLSR, which involves clustering data and applying local regression models to improve predictions.

Limitations

The method may require a good initial global model for effective clustering and prediction.

Digital Object Identifier (DOI)

10.1186/1752-0509-5-90

Want to read the original?

Access the complete publication on the publisher's website

View Original Publication