Student or educator? Sign up with your school email and get 6 months of Pro free.
CoChat

The DOI matches its Crossref record; 26 of the 26 references it cites were found in a registry or an index.

Optimal experiment selection for parameter estimation in biological differential equation models

Open access
Transtrum, Mark K, Qiu, Peng · 2012 · BMC Bioinformatics

Abstract

BACKGROUND: Parameter estimation in biological models is a common yet challenging problem. In this work we explore the problem for gene regulatory networks modeled by differential equations with unknown parameters, such as decay rates, reaction rates, Michaelis-Menten constants, and Hill coefficients. We explore the question to what extent parameters can be efficiently estimated by appropriate experimental selection. RESULTS: A minimization formulation is used to find the parameter values that best fit the experiment data. When the data is insufficient, the minimization problem often has many local minima that fit the data reasonably well. We show that selecting a new experiment based on the local Fisher Information of one local minimum generates additional data that allows one to successfully discriminate among the many local minima. The parameters can be estimated to high accuracy by iteratively performing minimization and experiment selection. We show that the experiment choices are roughly independent of which local minima is used to calculate the local Fisher Information. CONCLUSIONS: We show that by an appropriate choice of experiments, one can, in principle, efficiently and accurately estimate all the parameters of gene regulatory network. In addition, we demonstrate that appropriate experiment selection can also allow one to restrict model predictions without constraining the parameters using many fewer experiments. We suggest that predicting model behaviors and inferring parameters represent two different approaches to model calibration with different requirements on data and experimental cost.

Overview

The claims and statistics of this paper haven’t been checked yet.

Free: it reads the open-access text and checks the first 5 claims that cite a source we can check. Create a free account to check every claim.

A few free checks a day · papers already verified open instantly · results are public

Integrity signals

Registry

Crossref · 10.1186/1471-2105-13-181

Cites

26 of 26 references verified0 citation errors · 0 not checkable

Statistics

Figures

Figure 1 A graphical representation of model 1. Six genes are represented by network nodes. The network has eight edges, representing the protein mediated gene interactions.
Figure 1 A graphical representation of model 1. Six genes are represented by network nodes. The network has eight edges, representing the protein mediated gene interactions.CC BY · PubMed Central
Figure 2 Convergence criterion for minimization algorithm. The set of model predictions in data space form a surface known as the model manifold. By comparing the angle between the tangent plane to the Model Manifold and the vector of unfit data, we can monitor convergence to a local minimum of the cost. If αis approximately a right angle, then we are likely near a local minimum. Fits separated by more than |r(θ)|cosα are likely to correspond to different minima.
Figure 2 Convergence criterion for minimization algorithm. The set of model predictions in data space form a surface known as the model manifold. By comparing the angle between the tangent plane to the Model Manifold and the vector of unfit data, we can monitor convergence to a local minimum of the cost. If αis approximately a right angle, then we are likely near a local minimum. Fits separated by more than |r(θ)|cosα are likely to correspond to different minima.CC BY · PubMed Central
Figure 3 Cost barrier between two minima. Although we identified several distinct local minima with that were able to fit the data well, a direct line path connecting these minima reveal that the cost barrier between them is very small. Here we see two points for which the cost barrier between them is about 50. Most randomly selected parameter values have a cost between 1,000 and 10,000, illustrating that this barrier is relatively small. It is possible that these minima are not distinct but are connected by a winding canyon.
Figure 3 Cost barrier between two minima. Although we identified several distinct local minima with that were able to fit the data well, a direct line path connecting these minima reveal that the cost barrier between them is very small. Here we see two points for which the cost barrier between them is about 50. Most randomly selected parameter values have a cost between 1,000 and 10,000, illustrating that this barrier is relatively small. It is possible that these minima are not distinct but are connected by a winding canyon.CC BY · PubMed Central
Figure 4 Paths for distinct minima as data is collected. The path through parameter space (projected onto the first and second principal component of the ensemble of initial fits) as data is added to minimize the local variance. Although each path corresponds to a different sequence of experiments, they all arrive at the same estimate of the final parameters.
Figure 4 Paths for distinct minima as data is collected. The path through parameter space (projected onto the first and second principal component of the ensemble of initial fits) as data is added to minimize the local variance. Although each path corresponds to a different sequence of experiments, they all arrive at the same estimate of the final parameters.CC BY · PubMed Central
Figure 5 Confidence interval reduction by experiment selection. With the initial data, there is a large region of acceptable fits (black) which encompasses all the good fits found in the initial search. Adding a new experiment reduces the confidence region (red) and the best fit moves to a new point within the new confidence region. A second new experiment reduces it further (blue). This process is iterated until the confidence region is acceptably small.
Figure 5 Confidence interval reduction by experiment selection. With the initial data, there is a large region of acceptable fits (black) which encompasses all the good fits found in the initial search. Adding a new experiment reduces the confidence region (red) and the best fit moves to a new point within the new confidence region. A second new experiment reduces it further (blue). This process is iterated until the confidence region is acceptably small.CC BY · PubMed Central
Figure 6 Independent model degrees of freedom. Sample data from model 1 for the mRNA and protein concentration corresponding to gene 2 as well as the two genes which regulate these concentrations (genes 1 and 6). By noting that gene 2 is promoted by gene 1 and inhibited by gene 6, it is clear that these time series contain redundant information. The rise and subsequent fall in the mRNA and protein concentrations of gene 2 can be predicted from the time series of genes 1 and 6. Measuring all six time series would be less effective than alternative experiments that probed independent degrees of freedom.
Figure 6 Independent model degrees of freedom. Sample data from model 1 for the mRNA and protein concentration corresponding to gene 2 as well as the two genes which regulate these concentrations (genes 1 and 6). By noting that gene 2 is promoted by gene 1 and inhibited by gene 6, it is clear that these time series contain redundant information. The rise and subsequent fall in the mRNA and protein concentrations of gene 2 can be predicted from the time series of genes 1 and 6. Measuring all six time series would be less effective than alternative experiments that probed independent degrees of freedom.CC BY · PubMed Central

Verification

Claims the paper makes

The claims and statistics of this paper haven’t been checked yet.

Free: it reads the open-access text and checks the first 5 claims that cite a source we can check. Create a free account to check every claim.

A few free checks a day · papers already verified open instantly · results are public

Claims other papers make about this one

Each is checked against this paper’s own text: does it say what the citing paper says it does?

Self-check

Percentages, sums and reported differences re-computed from the paper’s own numbers.

The self-check needs the paper’s full text.

Citations

Cites

Show all 26 references

Research works printed but not linked: 1

  1. Brown KS: Signal Transduction, Sloppy Models, and Statistical Mechanics. PhD thesis. Cornell University 2003-

Cited by

See all 34

Metadata

Authors
Transtrum, M. K., & Qiu, P.
Affiliations
The University of Texas MD Anderson Cancer Center (US)
Corresponding author
Qiu, Peng
Venue
BMC Bioinformatics
Published
2012-07-27
Volume
13
Issue
1
Pages
181-181
ISSN
1471-2105 (ISSN-L)
OpenAlex
W2100560953
Cited by
34
References
26
Citations by year
2026: 3 · 2025: 2 · 2023: 2 · 2022: 5 · 2021: 1 · 2019: 1 · 2018: 6 · 2017: 2 · 2016: 4 · 2015: 4 · 2014: 2 · 2013: 2
FWCI
1.03
Type
journal article
Language
en
Open access
Licence
  • cc-by
  • creativecommons.org/licenses/by/2.0 (tdm)
Locations
  • BMC Bioinformatics (publishedVersion, cc-by)
  • PubMed Central (publishedVersion, cc-by)
  • PubMed (publishedVersion)
  • DOAJ (DOAJ: Directory of Open Access Journals) (submittedVersion, cc-by-sa)
Funders
  • University of Texas MD Anderson Cancer Center ((CA016672), P30-CA-016672)
  • National Cancer Institute (U24 CA143883 02 S1, U24CA143883, P30CA016672)
Topics
  • Life Sciences › Biochemistry, Genetics and Molecular Biology › Molecular Biology › Gene Regulatory Network Analysis
  • Physical Sciences › Engineering › Control and Systems Engineering › Control Systems and Identification
  • Social Sciences › Decision Sciences › Management Science and Operations Research › Optimal Experimental Design Methods
Keywords
parameter estimation, gene regulatory networks, differential equation model, local minima, model calibration, Michaelis-Menten constant, Hill coefficient, model discrimination, model reduction
MeSH terms
Algorithms, Mathematical Computing, Models, Genetic, Research Design, Gene Regulatory Networks

About these checks

Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.

Statistics and arithmetic are recomputed, not judged.

Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.

What counts as a claim ›How we measure confidence ›How we check this ›

Are you an author, or do you see a problem with this page? Request a review

How to cite

Cite CoChat for the checks on this page, and the paper for its own findings.

This page
CoChat. (2026). Citation check: Optimal experiment selection for parameter estimation in biological differential equation models. https://app.cochat.ai/papers/10.1186/1471-2105-13-181
The paper
Transtrum, M. K., & Qiu, P. (2012). Optimal experiment selection for parameter estimation in biological differential equation models. BMC Bioinformatics. https://doi.org/10.1186/1471-2105-13-181

Data sources

Bibliographic data: Semantic Scholar, OpenAlex, Crossref.

Contains information from the Semantic Scholar Academic Graph Datasets, made available under the ODC Attribution License (ODC-BY 1.0).

Semantic Scholar’s data is described in: Kinney, R., et al. (2023). The Semantic Scholar Open Data Platform. arXiv:2301.10140.