Jul 2026· Annual Conference on Genetic and Evolutionary Computation· pp. 1056-1066· 0 citations· 71 references
Computer Science
Abstract
The vital functions of life are driven by proteins — long folded strings of amino acids. Millions of proteins exist in nature, all optimized for a certain set of operating conditions. One particularly important trait of a protein is thermostability, or the tendency to maintain form and function at high temperatures. Protein thermostability is important for both biological organisms and industrial enzymatic processes that utilize them. This work proposes a novel hybrid EC approach, employing the EAs Simulating Molecular Evolution (EASME) software, for evolving thermostable variants of proteins with minimal computational overhead and customized problem-specific fitness functions. A variation of NSGA-II with a novel recombination/mutation interplay was designed to tackle the specific challenges of working with proteins. The results of this work, which computationally improved the thermostability of a protein more effectively than a state of the art machine learning model from 2025, demonstrate that EC may have been historically undervalued in the realm of protein design. The pipeline designed for this work could be applied to numerous other proteins, easily expanded to optimize other traits of those proteins, or employed on another problem entirely where mutations have a notable fitness cost.
Engineering biology is a critical technology of global significance, which through the application of rigorous engineering principles promotes the industrialization of biology. A major roadblock towards the sustainable impact of this technology remains in the lack of confidence in the predictability and reproducibility of biological design. This sets new requirements for the development of underpinning metrology which will enable the benchmark assessment of bioengineered systems and processes.
The understanding of the principles governing the folding of primary amino acid sequences has allowed the engineering of de novo peptide and proteins with desirable functions, including self-assembling virus-like particles (VLPs). More recently, machine learning algorithms have arisen as powerful tools to accurately predict the 3D protein structures. The training of these models with larger, high-fidelity datasets opens the possibility to the AI-assisted design of peptides and proteins with promising applications in fields such as cell and gene therapy or vaccine and drug development.
The reproducibility of the emerging manufacturing processes and the traceability of the materials’ desired properties is essential to ensure their efficacy and safety. De novo peptide and protein standards of well-characterized identity, purity, structure and activity are therefore needed to benchmark AI-driven engineered proteins. The CCQM Protein Analysis Working Group is running a series of interlaboratory comparisons to assess National Metrology Institutes’ (NMIs) capabilities to characterize peptide and protein pure standards materials. The chemistry department of the BIPM and the National Physics Laboratory (UK) have collaborated to investigate potential candidate materials for such comparisons, including the measurement of the purity of a de novo peptide, C3triskelion, capable of self-assembling into artificial virus-like capsids exhibiting strong antimicrobial activity. The material could also be a candidate reference material for VLPs, such as gene-delivery products.
The methodology developed at the BIPM to assign the purity of the C3triskelion included the mass balance method, qNMR and amino acid analysis. Despite challenges in the determination of structurally related impurities, the applied methods showed consistent results, demonstrating for the first time the possibility to value assign the mass fraction content with well-defined measurement uncertainty for this type of bioengineered material.
In addition to describing the measurement methods that have been developed for materials such as triskelion, the poster will also present potential future candidate materials for comparisons and their potential applications, notably how they may be employed to: confirm the purity of a commercial VLP, or virus-derived product as required by the manufacturer or a regulatory body; assist in the quantification of the encapsulation efficacy of a designed gene-delivery system; support validation of VLP performance in different sample matrices, in vitro, and in cell extracts and ultimately in live cells and tissues; and provide a route for measuring the amount of a desired material in target media with well-defined uncertainty traceable to well characterized reference materials.
G. Martos, Andrea Briones, R. Josephs et al.· 150th anniversary of the Met...· 0 citations
Results indicate that supervised machine learning can help guide the construction of high-value enzyme libraries with expanded catalytic scope, and suggest that supervised machine learning can help guide the construction of high-value enzyme libraries with expanded catalytic scope.
Ravi G. Lal, Jason Yang, Ziyan Zhang et al.· bioRxiv· 0 citations
ThermoFusion, a hybrid deep learning framework that integrates 3D protein structure embeddings from ThermoMPNN with sequence-based embeddings from the pretrained protein language model ESM2 to predict the effects of single-point mutations on protein stability is presented.
Yao Wei, I. Eberini, Fabian Meyer· bioRxiv· 0 citations
This review examines enzyme engineering from classical methods to AI-assisted biocatalyst development, highlighting key advances, challenges, and emerging trends in autonomous laboratories, sustainable biocatalysis, and computational protein design.
Mati Ullah, Muhammad Rizwan, Vivian Andoh et al.· Journal of Agricultural and...· 0 citations
Protein dynamism and evolvability are key parameters linking enzyme flexibility to adaptive potential, yet how these concepts apply to nucleic acid enzymes remains largely unexplored. Here, we propose that threose nucleic acid (TNA), a genetic polymer that is more conformationally restricted than DNA, may be evolutionarily constrained by its preorganized backbone. RNA-cleavage profiles comparing the well-known 10-23 DNA enzyme (DNAzyme) with two in vitro selected TNA enzymes (threozymes) reveal striking differences in their temperature dependence. While the DNAzyme catalyzed reaction is optimal at 37 °C and weakly active at 50 °C, threozymes display the opposite trend, suggesting that TNA catalysis must overcome a higher free-energy barrier than DNA catalysis. Consistent with this view, the reaction becomes less temperature dependent for threozymes that operate with more flexible active sites. Together, these findings highlight the importance of backbone structure as a critical parameter for evolvability with implications for RNA world models and biomedical applications.
Erica Kim Man Lee, John C. Chaput· Nature Communications· 0 citations
In this review, a review of recent in vivo hypermutation tools that enable rapid sampling of the vast evolutionary landscape, all while supporting simultaneous selection of the best proteins within living organisms are discussed.