“…When the aim is only the prediction of the final folded structure (PSP), the strategy to tackle the deceptiveness of the energy model is to try to provide conformational models (decoys) that minimize energy but which, at the same time, present a diversified structural distribution. The use of niching methods in evolutionary computation is a straightforward means of addressing the problem [44], since these niching methods enforce the distribution of the population (in this case of PSP, possible protein conformations) in different areas or niches of the energy landscape corresponding to different energy minima (and possibly to structural variants close to the native structure). However, this is not the objective in the present study, since the ANN only uses information from the (imperfect) Rosetta energy landscape to fold the structures towards lowenergy areas.…”
Protein folding is the dynamic process by which a protein folds into its final native structure. This is different to the traditional problem of the prediction of the final protein structure, since it requires a modeling of how protein components interact over time to obtain the final folded structure. In this study we test whether a model of the folding process can be obtained exclusively through machine learning. To this end, protein folding is considered as an emergent process and the cellular automata tool is used to model the folding process. A neural cellular automaton is defined, using a connectionist model that acts as a cellular automaton through the protein chain to define the dynamic folding. Differential evolution is used to automatically obtain the optimized neural cellular automata that provide protein folding. We tested the methods with the Rosetta coarse-grained atomic model of protein representation, using different proteins to analyze the modeling of folding and the structure refinement that the modeling can provide, showing the potential advantages that such methods offer, but also difficulties that arise.
“…When the aim is only the prediction of the final folded structure (PSP), the strategy to tackle the deceptiveness of the energy model is to try to provide conformational models (decoys) that minimize energy but which, at the same time, present a diversified structural distribution. The use of niching methods in evolutionary computation is a straightforward means of addressing the problem [44], since these niching methods enforce the distribution of the population (in this case of PSP, possible protein conformations) in different areas or niches of the energy landscape corresponding to different energy minima (and possibly to structural variants close to the native structure). However, this is not the objective in the present study, since the ANN only uses information from the (imperfect) Rosetta energy landscape to fold the structures towards lowenergy areas.…”
Protein folding is the dynamic process by which a protein folds into its final native structure. This is different to the traditional problem of the prediction of the final protein structure, since it requires a modeling of how protein components interact over time to obtain the final folded structure. In this study we test whether a model of the folding process can be obtained exclusively through machine learning. To this end, protein folding is considered as an emergent process and the cellular automata tool is used to model the folding process. A neural cellular automaton is defined, using a connectionist model that acts as a cellular automaton through the protein chain to define the dynamic folding. Differential evolution is used to automatically obtain the optimized neural cellular automata that provide protein folding. We tested the methods with the Rosetta coarse-grained atomic model of protein representation, using different proteins to analyze the modeling of folding and the structure refinement that the modeling can provide, showing the potential advantages that such methods offer, but also difficulties that arise.
“…The memetic search follows a three-stage evolutionary process, as the fitness of the encoded conformations corresponds to different Rosetta score functions in each stage, while the first stage of Rosetta is used to define the initial population (with partially folded and different conformations). The memetic version is detailed in Varela and Santos ( 2019 , 2020 ). HybridDE outperforms the Rosetta ab initio protocol in obtaining conformations with minimum energy and under the same number of conformational energy evaluations.…”
Section: Methodsmentioning
confidence: 99%
“…This increases the chances of obtaining candidate structures close to the native structure. The crowding niching method was found to be the most useful niching technique in the application, given its simple parameter decision process, defining the CrowdingDE version, detailed in Varela and Santos ( 2020 , 2022 ). Fig.…”
Section: Methodsmentioning
confidence: 99%
“… 1953 ) to decide whether a structural fragment replaces a part of the current conformational structure of the target protein, with the goal of finding the structure with the minimum energy. Working with the fragment replacement technique as a local search operator, our evolutionary/memetic computational solutions ( HybridDE and CrowdingDE ) (Varela and Santos 2019 , 2020 , 2022 ) outperform Rosetta ab initio protocol when searching for structures with minimal energy and under the same number of conformational energy evaluations.…”
In this paper we discuss the advantages and problems of two alternatives for ab initio protein structure prediction. On one hand, recent approaches based on deep learning, which have significantly improved prediction results for a wide variety of proteins, are discussed. On the other hand, methods based on protein conformational energy minimization and with different search strategies are analyzed. In this latter case, our methods based on a memetic combination between differential evolution and the fragment replacement technique are included, incorporating also the possibility of niching in the evolutionary search. Different proteins have been used to analyze the pros and cons in both approaches, proposing possibilities of integration of both alternatives.
“…In this line, our previous research used memetic algorithms (MAs) with protein atomic models, using the protein representation and energy model of Rosetta [7] (one of the most widely used software environments in PSP and protein design). Our HybridDE MA [9] combines the global search of Differential Evolution [6] with the local search provided by the protein fragment replacement technique, where the latter can locally refine protein structures maintained in the genetic population. Furthermore, given the inaccuracies of the Rosetta energy model, which provides a deceptive energy landscape in which the energy minimum need not correspond to the native structure, the crowding niching method was integrated into the MA (CrowdingDE) [9][10].…”
This paper presents a discussion of alternatives for protein structure prediction, showing the advantages and problems of recent alternatives based on deep learning and approaches based on energy optimization of protein solutions. A SARS-CoV-2 protein is included to exemplify the results.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.
