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

Abstract

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

Predicting novel mosquito-associated viruses from metatranscriptomic dark matter

“…This performance surpasses the well-known challenge of classifying short, non specific, and low informational content sequences ( 32 , 33 ). Notably, our approach of combining multiple predictive models, rather than relying on single models, further enhanced accuracy for shorter fragments, as indicated by previous studies ( 34 , 35 ). In our classification scenarios, the FP results (false mosquito-associated viruses and false arboviruses) pose a greater concern than those of FN.…”

“…To seamlessly transition from individual algorithm evaluations to collective decision-making, we leveraged the best-performing algorithms identified during 50 random resamplings in a soft voting strategy. This mechanism aggregates predictions from diverse machine learning models, improving robustness and accuracy ( 34 , 35 ). This approach enhances the robustness and reliability of the overall classification process by leveraging the strengths of each contributing model to provide comprehensive classification.…”

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

“…A typical production process is shown in Figure 1a. Process steps that are influenced by the epitope insertion are, for example, precipitation, where a varying amount of ammonium sulfate is required for VLP precipitation (Hillebrandt et al, 2020; Vormittag et al, 2020b), and VLP reassembly, which has been proposed to depend on VLP zeta potential (Rüdt et al, 2019). VLPs are disassembled (dissociated into capsomeres) and reassembled (capsomeres triggered to form capsids) to improve structural homogeneity, stability, and immunogenicity (Zhao, Allen, et al, 2012; Zhao, Modis, et al, 2012).…”

Size-selective downstream processing of virus particles and non-enveloped virus-like particles

“…Examples are the precipitation of norovirus VLPs using PEG ( Koho et al, 2012 ) or precipitation of different HBcAg VLP candidates ( Schumacher et al, 2018 ; Zhang et al, 2021 ) and adeno-associated virus strains using ammonium sulfate ( Tomono et al, 2016 ; 2018 ). For the screening of VLP precipitation conditions, advanced analytical and predictive tools were recently developed ( Vormittag et al, 2020 ; Wegner & Hubbuch, 2022 ). Integration of precipitation, wash, and re-dissolution with cross-flow filtration led to synergies improving process performance ( Hillebrandt et al, 2020 ).…”