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.
Abstract
Protein thermostability is a critical property for both industrial and biomedical enzyme applications, yet experimental evaluation of mutation-induced stability changes remains laborious and costly. Here, we present 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 (ΔΔG). ThermoFusion exhibits robust generalization, maintaining high predictive accuracy across out of distribution sequences with low identity to the training set – a scenario where many other machine learning models, including ThermoMPNN and state-of-the-art tools, perform poorly due to reliance on memorization. Benchmarking on a curated enzyme dataset comprising of 105 enzymes and 3144 mutations shows that ThermoFusion reliably identifies stabilizing mutations while accurately predicting stability for enzymes beyond its training set. These results establish ThermoFusion as a powerful tool for rational enzyme design beyond its training set.
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.
J. S. Browning, D. Tauritz, John Beckmann· Annual Conference on Genetic...· 0 citations
Predicting protein stability, like changes in melting temperature (ΔTm) caused by mutations, is a critical task in therapeutic protein engineering and drug discovery. This is reflected by a growing solution space, including both AI-based sequence and structure based methods. This paper demonstrates that accurate ΔTm prediction does not require structural input features, but can achieve state-of-the-art results with a careful training design for large sequence-based protein language models. We combine an autoresearch-inspired setup search with controlled ablation studies and show that a well-tuned sequence-only ESM2-650M model [6] outperforms structure-informed methods in our benchmark, achieving the lowest error (MAE/RMSE) and competitive Pearson correlation without pH or structural inputs. We further show that choices such as loss function, pooling strategy, auxiliary supervision, and finetuning regime materially affect performance.
Daniel Siegismund, Mario Wieser, E. Natali et al.· bioRxiv· 0 citations
A systematic NMR-characterized dataset of mutants of the GA/GB model fold-switching system is presented and it is found that this benchmark revealed variable and position-dependent performance across methods, with certain AlphaFold2-based algorithms able to predict mutant effects at individual sites, indicating some understanding of physical effects of residue substitutions.
Nathaniel R. Felbinger, K. Carillo, Yihong Chen et al.· bioRxiv· 0 citations
UniStab is introduced, an end-to-end framework for predicting stability changes across all mutation types by leveraging the implicit geometric reasoning of a pre-trained folding model and demonstrates state-of-the-art performance, particularly in the challenging scenarios of multi-point mutations and indels.
Hong Tan, Shenggeng Lin, Yi Xiong· Chemical Science· 0 citations
A small-sample, accelerated evolution strategy that integrates focused rational iterative site-specific mutagenesis (FRISM) with the EVOLVEpro model is reported, providing a robust, "lightweight" machine learning framework for the rapid development of new-to-nature photoenzymatic transformations.
Four recently released ES and ER prediction models are benchmarked and it is suggested that interaction-aware representations from full biomolecular complexes may provide a promising basis for enzyme prioritization.
Elizabeth H. Mahood, N. Komorníková, Tom'avs Pluskal et al.· 0 citations