Ensembles of Hydrophobicity Scales as Potent Classifiers for Chimeric Virus-Like Particle Solubility – An Amino Acid Sequence-Based Machine Learning Approach

Vormittag, Philipp; Klamp, Thorsten; Hubbuch, Jürgen

doi:10.3389/fbioe.2020.00395

Cited by 8 publications

(18 citation statements)

References 71 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The model is, at this stage, incapable of describing this different behavior. As previously suggested, this could be related to 3-D phenomena that cannot be described by a sequence-based approach (Vormittag et al, 2020). Knowledge of the abovedescribed systematic misclassification helps (a) to conclude that this insertion strategy may be disadvantageous with respect to solubility, and (b) to adjust the model, so that the model's blind spot is compensated.…”

Section: Introductionmentioning

confidence: 91%

“…None of these individual attributes describe the cVLP solubility landscape comprehensively. This is underlined by findings, in which combining different attributes improved the solubility model's performance (Vormittag et al, 2020). Each amino acid makes a unique contribution to protein solubility, e.g., based on its charge, volume, or specific interactions.…”

Section: Introductionmentioning

confidence: 99%

“…While this flexibility of antigenic display is one of the major advantages of VLPs (Pumpens et al, 2008), recombinant insertion of epitopes often results in expression of insoluble structures (Karpenko et al, 2000;Billaud et al, 2005). Factors affecting cVLP solubility have been described as, for example, insert charge (Whitacre et al, 2009), amino acid side chain volume (Karpenko et al, 2000), or the content of specific amino acids, such as tryptophan or arginine (Vormittag et al, 2020). None of these individual attributes describe the cVLP solubility landscape comprehensively.…”

Section: Introductionmentioning

confidence: 99%

“…In our previous study, we demonstrated the potential of optimizing the model's prediction based on the contingency matrices of the individual insertion strategies (Vormittag et al, 2020). The dataset used consisted of 568 chimeric HBcAg constructs, created by a grid of 71 different inserts and eight insertion strategies.…”

Section: Introductionmentioning

confidence: 99%

“…The different strategies are meant to optimize the integration of the foreign epitope into the VLP sequence and would ideally result in an integration that produces a soluble cVLP. Analyzing the strategies showed that the model systematically overestimated or underestimated certain insertion strategies with respect to the predicted solubility (Vormittag et al, 2020). To recapitulate briefly, a strategy that is overrepresented in the training false-positive (FP) group has overestimated solubility in relation to the other strategies.…”

Section: Introductionmentioning

confidence: 99%

See 4 more Smart Citations

Optimization of a Soft Ensemble Vote Classifier for the Prediction of Chimeric Virus-Like Particle Solubility and Other Biophysical Properties

Vormittag

Klamp

Hubbuch

2020

Front. Bioeng. Biotechnol.

Self Cite

View full text Add to dashboard Cite

Chimeric virus-like particles (cVLPs) are protein-based nanostructures applied as investigational vaccines against infectious diseases, cancer, and immunological disorders. Low solubility of cVLP vaccine candidates is a challenge that can prevent development of these very substances. Solubility of cVLPs is typically assessed empirically, leading to high time and material requirements. Prediction of cVLP solubility in silico can aid in reducing this effort. Protein aggregation by hydrophobic interaction is an important factor driving protein insolubility. In this article, a recently developed soft ensemble vote classifier (sEVC) for the prediction of cVLP solubility was used based on 91 literature amino acid hydrophobicity scales. Optimization algorithms were developed to boost model performance, and the model was redesigned as a regression tool for ammonium sulfate concentration required for cVLP precipitation. The present dataset consists of 568 cVLPs, created by insertion of 71 different peptide sequences using eight different insertion strategies. Two optimization algorithms were developed that (I) modified the sEVC with regard to systematic misclassification based on the different insertion strategies, and (II) modified the amino acid hydrophobicity scale tables to improve classification. The second algorithm was additionally used to synthesize scales from random vectors. Compared to the unmodified model, Matthew's Correlation Coefficient (MCC), and accuracy of the test set predictions could be elevated from 0.63 and 0.81 to 0.77 and 0.88, respectively, for the best models. This improved performance compared to literature scales was suggested to be due to a decreased correlation between synthesized scales. In these, tryptophan was identified as the most hydrophobic amino acid, i.e., the amino acid most problematic for cVLP solubility, supported by previous literature findings. As a case study, the sEVC was redesigned as a regression tool and applied to determine ammonium sulfate concentrations for the

show abstract

Section: Introductionmentioning

confidence: 91%

Section: Introductionmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 3 more Smart Citations

Optimization of a Soft Ensemble Vote Classifier for the Prediction of Chimeric Virus-Like Particle Solubility and Other Biophysical Properties

Vormittag

Klamp

Hubbuch

2020

Front. Bioeng. Biotechnol.

Self Cite

View full text Add to dashboard Cite

show abstract

Engineering Tobacco Mosaic Virus and Its Virus‐Like‐Particles for Synthesis of Biotemplated Nanomaterials

Lee

Pussepitiyalage

Lee

et al. 2020

Biotechnology Journal

View full text Add to dashboard Cite

Biomolecules are increasingly attractive templates for the synthesis of functional nanomaterials. Chief among them is the plant tobacco mosaic virus (TMV) due to its high aspect ratio, narrow size distribution, diverse biochemical functionalities presented on the surface, and compatibility with a number of chemical conjugations. These properties are also easily manipulated by genetic modification to enable the synthesis of a range of metallic and non-metallic nanomaterials for diverse applications. This article reviews the characteristics of TMV and related viruses, and their virus-like particle (VLP) derivatives, and how these may be manipulated to extend their use and function. A focus of recent efforts has been on greater understanding and control of the self-assembly processes that drive biotemplate formation. How these features have been exploited in engineering applications such as, sensing, catalysis, and energy storage are briefly outlined. While control of VLP surface features is well-established, fewer tools exist to control VLP self-assembly, which limits efforts to control template uniformity and synthesis of certain templated nanomaterials. However, emerging advances in synthetic biology, machine learning, and other fields promise to accelerate efforts to control template uniformity and nanomaterial synthesis enabling more widescale industrial use of VLP-based biotemplates.

show abstract

Process development for cross‐flow diafiltration‐based VLP disassembly: A novel high‐throughput screening approach

Hillebrandt

Vormittag

Dietrich

et al. 2021

Biotech & Bioengineering

Self Cite

View full text Add to dashboard Cite

Virus‐like particles (VLPs) are particulate structures, which are applied as vaccines or delivery vehicles. VLPs assemble from subunits, named capsomeres, composed of recombinantly expressed viral structural proteins. During downstream processing, in vivo‐assembled VLPs are typically dis‐ and reassembled to remove encapsulated impurities and to improve particle morphology. Disassembly is achieved in a high‐pH solution and by the addition of a denaturant or reducing agent. The optimal disassembly conditions depend on the VLP amino acid sequence and structure, thus requiring material‐consuming disassembly experiments. To this end, we developed a low‐volume and high‐resolution disassembly screening that provides time‐resolved insight into the VLP disassembly progress. In this study, two variants of C‐terminally truncated hepatitis B core antigen were investigated showing different disassembly behaviors. For both VLPs, the best capsomere yield was achieved at moderately high urea concentration and pH. Nonetheless, their disassembly behaviors differed particularly with respect to disassembly rate and aggregation. Based on the high‐throughput screening results, a diafiltration‐based disassembly process step was developed. Compared with mixing‐based disassembly, it resulted in higher yields of up to 0.84 and allowed for integrated purification. This process step was embedded in a filtration‐based process sequence of disassembly, capsomere separation, and reassembly, considerably reducing high‐molecular‐weight species.

show abstract

Ensembles of Hydrophobicity Scales as Potent Classifiers for Chimeric Virus-Like Particle Solubility – An Amino Acid Sequence-Based Machine Learning Approach

Cited by 8 publications

References 71 publications

Optimization of a Soft Ensemble Vote Classifier for the Prediction of Chimeric Virus-Like Particle Solubility and Other Biophysical Properties

Optimization of a Soft Ensemble Vote Classifier for the Prediction of Chimeric Virus-Like Particle Solubility and Other Biophysical Properties

Engineering Tobacco Mosaic Virus and Its Virus‐Like‐Particles for Synthesis of Biotemplated Nanomaterials

Process development for cross‐flow diafiltration‐based VLP disassembly: A novel high‐throughput screening approach

Contact Info

Product

Resources

About