The hybrid experimental simplex algorithm – An alternative method for ‘sweet spot’ identification in early bioprocess development: Case studies in ion exchange chromatography

Abstract

The capacity to locate efficiently a subset of experimental conditions necessary for the identification of an operating envelope is a key objective in many studies. We have shown previously how this can be performed by using the simplex algorithm and this paper now extends the approach by augmenting the established simplex method to form a novel hybrid experimental simplex algorithm (HESA) for identifying 'sweet spots' during scouting development studies. The paper describes the new algorithm and illustrates its use in two bioprocessing case studies conducted in a 96-well filter plate format. The first investigates the effect of pH and salt concentration on the binding of green fluorescent protein, isolated from Escherichia coli homogenate, to a weak anion exchange resin and the second examines the impact of salt concentration, pH and initial feed concentration upon the binding capacities of a FAb', isolated from E. coli lysate, to a strong cation exchange resin. Compared with the established algorithm, HESA was better at delivering valuable information regarding the size, shape and location of operating 'sweet spots' that could then be further investigated and optimized with follow up studies. To test how favorably these features of HESA compared with conventional DoE (design of experiments) methods, HESA results were also compared with approaches including response surface modeling experimental designs. The results show that HESA can return 'sweet spots' that are equivalently or better defined than those obtained from DoE approaches. At the same time the deployment of HESA to identify bioprocess-relevant operating boundaries was accompanied by comparable experimental costs to those of DoE methods. HESA is therefore a viable and valuable alternative route for identifying 'sweet spots' during scouting studies in bioprocess development.

An Intelligent Automation Platform for Rapid Bioprocess Design

“…15 Starting with an initial experiment design, the strategy is to sequentially revise the parameter estimates according to data generated from the previous experiments until the estimates are within a specified accuracy. 16–18 …”

“…15,17,18 However, both the above methods have the potential to be implemented in an automated fashion.…”

Hybrid optimization of preparative chromatography for a ternary monoclonal antibody mixture

“…Central to the hybrid approach is the idea of systematically taking stock of all the available data and increasing process insight as the optimization progresses. The simplex is a hill climbing method that navigates based on experimental data and finds optima with high likelihood and accuracy . The performance of the simplex depends on the path leading to the optimum and varies, for example, with the starting location, the distance from the optimum or constraints encountered with respect to operational parameters.…”

Simplex optimization: A tutorial approach and recent applications in analytical chemistry

“…Cai et al [44] have used an optimization method based on the modified simplex method with the fuzzy set theory in the optimization of chromatographic separation and derivation conditions of amino acids neurotransmitters. Konstantinidis et al [21] have applied a novel hybrid experimental simplex algorithm to identify "sweet spots" in bioprocess development using ion exchange chromatography.…”