Design of experiments for steady-state system identification with applications in genetic and business network modelinG

ABSTRACT

This paper focuses on the inverse problem of predicting inputs from measured outputs in the context of linear systems in steady-state. For system identification, we propose forward network identification regression (FNIR) and experimental planning involving simultaneously perturbing more than a single gene concentration using D-optimal designs. The proposed methods are compared with alternatives using simulation and data sets motivated by the SOS pathway for Escherichia coli bacteria. Findings include that the optimal experimental planning can likely improve the sensitivity, specificity, and efficiency of the process of deriving genetic networks. Topics for further research are also suggested in this paper.

KEYWORDS:

• Optimal Design of Experiments
• D-Optimality
• inverse Regression
• transcriptional Networks
• statistical Simulation

Acknowledgments

We thank Allen Miller for contributing significantly to the development of the random resilient matrices (RRM) simulation method. We thank James Collins and Timothy Gardner for sharing their MATLAB® code with us. Diego Di Bernardo shared a relevant preprint with us. Ning Zheng and Joseph Fiksel provided many helpful discussions and Mikhail Bernshteyn generated the designs. Andrew Ashman, Ravishankar Rajagopalan, and Anthony West provided important assistance with our nonlinear sensitivity analysis. The editor and anonymous reviewers provided valuable feedback which caused us to greatly improve the clarity of our discussion in general and about the linearity assumptions in particular.

Disclosure statement

No potential conflict of interest was reported by the authors.

Additional information

Notes on contributors

Cenny Taslim

Cenny Taslim, PhD. is a postdoctoral research in the Comprehensive Cancer Center at The Ohio State University Wexner Medical Center. Dr. Taslim has developed statistical methodology to analyze high-throughput next generation sequencing data. Her current research focuses on integrative analysis of “omics” data, epidemiologic and clinical data to understand the molecular mechanism of breast cancer risk. Dr. Taslim has authored 11 peer-reviewed publications.

Theodore T. Allen

Theodore T. Allen, Ph.D. is an associate professor of Integrated Systems Engineering at The Ohio State University. His interests are at the boundary of optimization, statistics, control theory, and machine learning. He has over 50 peer-reviewed publications.

Mario Lauria

Mario Lauria, Ph.D. is a researcher at the Microsoft Research - University of Trento Centre for Computational and Systems Biology. He is a former Fulbright scholar and is a Senior Member of the IEEE. He is the author of over 50 peer-reviewed publications.

Shih-Hsien Tseng

Shih-Hsien Tseng, Ph.D. is an Assistant Professor in the Department of Business Administration at Chung Yuan Christian University. His primary interests relate to new methods for optimal experimental design and novel statistical methods for quality management and decision-making. He has published in academic journals and worked at ACE Geosynthetics for about 3 years applying Six Sigma and related techniques to improve the manufacturing process. His e-mail address is [email protected].