Evolutionary experiments for self-assembling amphiphilic systems

Forlin, Michele; Poli, Irene; March, Davide De; Packard, Norman H.; Gazzola, Gianluca; Serra, Roberto

doi:10.1016/j.chemolab.2007.09.007

Cited by 18 publications

(9 citation statements)

References 14 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Models are commonly used in design of experiments (Caschera et al, 2010), but their use here to provide virtual experimental results and adaptively establish the trade‐off between exploration and exploitation is novel. Our use of Evo‐DoE is an enhanced version of the techniques used recently in a different optimization experiment (Forlin et al, 2008), and goes beyond the use of standard (Corma et al, 2005) or model‐assisted (Stein, 1999) genetic algorithms for evolutionary design of experiments. Our approach is similar to a family of iterative nonlinear optimization techniques that employ kriging models (Jones et al, 1998) to interpolate experimental observations (Cawse et al, 2010).…”

Section: Resultsmentioning

confidence: 99%

Coping with complexity: Machine learning optimization of cell‐free protein synthesis

Caschera

Bedau

Buchanan³

et al. 2011

Biotech & Bioengineering

View full text Add to dashboard Cite

Biological systems contain complex metabolic pathways with many nonlinearities and synergies that make them difficult to predict from first principles. Protein synthesis is a canonical example of such a pathway. Here we show how cell-free protein synthesis may be improved through a series of iterated high-throughput experiments guided by a machine-learning algorithm implementing a form of evolutionary design of experiments (Evo-DoE). The algorithm predicts fruitful experiments from statistical models of the previous experimental results, combined with stochastic exploration of the experimental space. The desired experimental response, or evolutionary fitness, was defined as the yield of the target product, and new experimental conditions were discovered to have ∼ 350% greater yield than the standard. An analysis of the best experimental conditions discovered indicates that there are two distinct classes of kinetics, thus showing how our evolutionary design of experiments is capable of significant innovation, as well as gradual improvement.

show abstract

Section: Resultsmentioning

confidence: 99%

Coping with complexity: Machine learning optimization of cell‐free protein synthesis

Caschera

Bedau

Buchanan³

et al. 2011

Biotech & Bioengineering

View full text Add to dashboard Cite

show abstract

“…2), we not only increased the experimental throughput and the procedural manageability, but we lowered the costs and reduced the amount of contributory factors. This reduction of dimensionality and the parallelism in vesicle formation provided by the novel vesicle formation method presented herein may prove to be useful in evolutionary design of experiments by shortening the search toward the optimality region of the search space [2]. The application of microtiter plates may further enable the automatization in vesicle formation.…”

Section: Resultsmentioning

confidence: 99%

“…In this context, PACE has created the foundation for a new generation of embedded IT to build evolvable complex information systems using programmable chemical systems that converge to artificial cells [1]. Because experiments were realized both in the laboratory and in simulation, findings of in vitro and in silico experiments interacted and lead to essential additions to the evolutionary approach in design of laboratory experimentation [2]. According to the guidelines of PACE, the creation of simple forms of life from scratch in the laboratory should have pursued a bottom-up strategy choosing simple organic compounds of low molecular weight over highly evolved polypeptides.…”

Section: Introductionmentioning

confidence: 99%

Towards Tailored Communication Networks in Assemblies of Artificial Cells

Hadorn

Burla

Hotz

2009

Artificial Life: Borrowing From Biology

View full text Add to dashboard Cite

“…Running the cross validation procedure, we obtain z best neural networks in terms of their CVRPE and their z designs whose responses are recorded. The ENN algorithm follows the bio‐inspired theory of Genetic Algorithms (GAs) , which is supported by a large number of successful applications, for example, . In GAs, a fitness function plays the role of score, and in our algorithm, it measures the goodness of the design.…”

Section: Detecting Optimal Designs From Observational Datamentioning

confidence: 99%

“…A different approach for efficiently designing experimental strategies is to adopt an evolutionary paradigm . This approach is fundamentally led by an optimization process that iteratively evolves the initial design toward some optimal regions of the experimental space or until a certain convergence criterion is satisfied.…”

Section: Introductionmentioning

confidence: 99%

T‐optimality and neural networks: a comparison of approaches for building experimental designs

Berni

March

Stefanini

2012

Appl Stoch Models Bus & Ind

View full text Add to dashboard Cite

This paper deals with optimal experimental design criteria and neural networks in the aim of building experimental designs from observational data. It addresses the following three main issues: (i) the introduction of two radically different approaches, namely T-optimal designs extended to Generalized Linear Models and Evolutionary Neural Networks Design; (ii) the proposal of two algorithms, based on model selection procedures, to exploit the information of already collected data; and (iii) the comparison of the suggested methods and corresponding algorithms by means of a simulated case study in the technological field.Results are compared by considering elements of the proposed algorithms, in terms of models and experimental design strategies. In particular, we highlight the algorithmic features, the performances of the approaches, the optimal solutions and the optimal levels of variables involved in a simulated foaming process. The optimal solutions obtained by the two proposed algorithms are very similar, nevertheless, the differences between the paths followed by the two algorithms to reach optimal values are substantial, as detailed step-by-step in the discussion.Optimal designs are basically model dependent; therefore, model discrimination has received great attention in literature also following different approaches, which can be grouped according to the methodological approach. The first class of contributions deals with model discrimination together with other issues, in particular the parameter estimation. In this regard, following the seminal contributions by [14,15] and [7], constrained optimality based on D-optimality is introduced, and the concept of canonical moments is expounded in order to find an optimal design and to choose the best model. In [16], an optimal robust design for polynomial regression models is developed by eliciting priors onˇ-coefficients as weight, as previously introduced in [15].In addition, in [17], generalization of previous optimality measures, such as in [4], is developed; while in [18], further issues are studied for the Fourier regression model and the power of the F -test. It must be noted that in [18], for discriminating among designs/models, the weights and the power of tests are evaluated by considering D-optimal restricted criterions.Undoubtedly, modeling discrimination and parameter estimation compared with T-optimality, as stated in [19], share common features and properties but, at the same time, the sequential nature of T-optimality may be viewed as a specific characteristic that is not a fundamental feature when applying D-optimality [1,5]. Nonetheless, as in [20] and [21], the T-optimality algorithm works through an adaptive sequential scheme, and the power of the F -test is affected by this selection process.A different approach for efficiently designing experimental strategies is to adopt an evolutionary paradigm [22][23][24][25][26][27]. This approach is fundamentally led by an optimization process that iteratively evolves the initial design toward some op...

show abstract

Evolutionary experiments for self-assembling amphiphilic systems

Cited by 18 publications

References 14 publications

Coping with complexity: Machine learning optimization of cell‐free protein synthesis

Coping with complexity: Machine learning optimization of cell‐free protein synthesis

Towards Tailored Communication Networks in Assemblies of Artificial Cells

T‐optimality and neural networks: a comparison of approaches for building experimental designs

Contact Info

Product

Resources

About