Using the 3D Beta-barrel Membrane Protein Predictor (3D-BMPP), a more accurate depiction of extended beta barrels and loops in regions outside the transmembrane domain is possible, with greater coverage of the overall structure. stemcells signals inhibitors In the context of protein analysis, 3DBMPP is a general method capable of handling protein families with limited sequences, and also proteins characterized by unique structural forms. Structure prediction of genome-wide MPs finds broad applicability in the use of 3DBMPP.

Protecting us from antigenic threats is the specific role of adaptive immunity. Within the framework of adaptive immunity, antibodies, as key effector proteins, are remarkable in their ability to recognize a virtually limitless scope of antigens. The antigen-binding region of antibodies, also known as the fragment variable (FV), is composed of two key parts: the framework region and the complementarity-determining regions. The framework (FR) is a combination of the light-chain framework (FRL) and the heavy-chain framework (FRH). In a similar vein, the complementarity-determining regions (CDRs) encompass light-chain CDRs 1 to 3 (CDRs L1-3) and heavy-chain CDRs 1 to 3 (CDRs H1-3). Across diverse antibody types, the Framework Regions (FRs) demonstrate a constant sequence and structure, yet the variations in sequences within the Complementarity Determining Regions (CDRs), which in turn influence the conformations of CDR loops, are the key to the different antigenic specificities. The preservation of structural elements within FR regions, coupled with the consistency of CDR conformations to a restricted repertoire of standard forms, enables the accurate prediction of antibody variable region models using homology modeling techniques. The amino acid sequence of an antibody can be used to predict its structure, facilitating the identification of antibody-antigen interface and enabling the modification of valuable therapeutic and biotechnological antibodies for enhanced affinity. Current homology modeling approaches for antibody variable regions, and the subsequent potential applications of these generated models, are detailed within this chapter.

Infectious SARS-CoV-2, a beta coronavirus, is responsible for the severe acute respiratory syndrome, COVID-19. The SARS-CoV-2 virus, demonstrating similarities to previous SARS and MERS strains, has infected a global population of nearly 650 million people, while the death toll has crossed six million by the end of December 2022. Using computational modeling approaches to understand viral proteins, this chapter investigates the different phases of viral lifecycles within hosts. This knowledge promises key insights into strategies for countering current and future viral threats. The recognition of these elements enables the effective determination of focal points in the search for novel drugs and vaccines.

Membrane transporter proteins, categorized into channels/pores and carriers, are crucial protein families with significant physiological and pharmacological implications. Membrane transporter proteins are the targets of presently used therapeutic compounds, like anti-arrhythmic, anesthetic, antidepressant, anxiolytic, and diuretic medications, which leads to their respective therapeutic actions. Human transporter three-dimensional structures’ absence impedes experimental research and drug development efforts. Homology modeling’s contribution to the development of structural models for membrane transporter proteins is reviewed within this chapter. Atomic-resolution structural templates, now more plentiful, combined with refined sequence alignment, secondary structure prediction, and model-building methodologies, and the significant advancement in computational resources, have collectively augmented the applicability of homology modeling in creating structural models of transporter proteins. Different facets of template selection, multiple-sequence alignment, model generation and optimization, model validation, and the application of transporter homology models in structure-based virtual ligand screening are addressed.

Protein regions, intrinsically disordered regions (IDRs), are characterized by a lack of fixed tertiary structure. Since these regions exhibit no ordered three-dimensional structures, they should be removed from the target selection process for homology modeling. The distinctive amino acid sequences of IDRs allow for their prediction, as they differ from the amino acid compositions of structured domains. This chapter offers a review of IDR prediction methods and demonstrates their application through a case study.

In the context of RNA data analysis, machine learning models find molecular representations to be extremely important. Intrinsically, useful molecular descriptors or fingerprints, showcasing the inherent structural and interactional data of RNA, can substantially raise the performance of all learning-based models. This research introduces two persistent models, persistent homology and persistent spectral, specifically tailored to RNA structure and interaction representations and their use in RNA data analyses. Persistent homology, unlike geometric and graph-based representations, is constructed using simplicial complexes, a higher-dimensional counterpart to graph models. The recent proposal of hypergraph-based embedded persistent homology represents a further generalization of simplicial complexes. In addition, molecular representations are proposed using persistent spectral models, which seamlessly blend filtration methods with spectral models encompassing spectral graphs, spectral simplicial complexes, and spectral hypergraphs. The persistent characteristics inherent in RNAs can be derived from these two persistent models and subsequently merged with machine learning models to delve into RNA structure, flexibility, dynamics, and function.

Understanding the structural consequences of a single amino acid substitution is paramount in the field of protein structural biology. This paper introduces recent progress in methods, permitting the decomposition of distortion effects into the substitution’s influence and the localized flexibility of the mutated position. A significant impediment lies in the necessity for a multitude of structures, each uniquely mutated. To bypass this impediment, we recommend utilizing molecular modeling tools; several software applications facilitate the creation of a model using a template, provided the sequence information. In substantiating our proof-of-concept, we utilize the human lysozyme, the gold standard. Both the wild-type and three mutant forms of the structure are cataloged in the PDB. Amyloid fibril formation is a consequence of two of these mutations, while the last one has no discernible effect. Finally, irrespective of the modeling algorithm, the side chain conformations at the mutation location are dependable, although the tools’ reach regarding long-range effects is restricted.

Olfactory receptors (ORs) are the predominant subfamily found within the class A G protein-coupled receptors (GPCRs). No empirical structural data on any OR has been obtained until now. To better understand the intricate interplay between receptor structure and function and to promote the discovery of ligands which can modulate receptor activity, homology modeling is proving to be a valuable strategy for proposing plausible OR models. For OR structure construction, this chapter details a general procedure, encompassing the acquisition of candidate templates, structure-based sequence alignment methods, 3D structural model creation, ligand docking simulations, and molecular dynamic simulations.

Membrane proteins, the G protein-coupled receptors (GPCRs), hold significant therapeutic importance. While experimental GPCR structures are proliferating, homology modeling retains significance in the study of these receptors and the discovery of new interacting molecules. This paper presents a comprehensive overview of current leading-edge methods for GPCR modeling, starting with template selection, progressing through refined sequence alignment and concluding with model refinement strategies.

Membrane proteins are difficult to characterize structurally through experimentation, currently represented in the Protein Data Bank by approximately 2% of the total structures. This difference in protein types, membrane versus soluble, dictates a smaller array of modeling techniques and a lower caliber of resultant models for membrane proteins. Recent improvements in protein expression, crystallization, and cryo-electron microscopy have led to an increased number of experimentally determined membrane protein structures, which can serve as blueprints for predicting the structures of comparable proteins via homology modeling techniques. Homology modeling, benefiting from a structural template, offers superior accuracy and simplicity compared to fold recognition or de novo modeling, methods typically resorted to when sequence similarity between the target sequence and structurally related proteins in databases dips below 25%. Homology modeling involves the mapping of a query sequence onto a single template structure, which is then refined. Due to the proliferation of templates, there’s frequently an overlap in the query sequence coverage among various templates. By employing multi-template modeling, the best template for each local segment can be pinpointed and incorporated into a singular model. Employing the Rosetta software suite, we detail a protocol for modeling membrane proteins using templates from diverse sources. RosettaScripts, RosettaCM, and RosettaMP, these integrated frameworks, along with the membrane scoring function, are exploited in this approach.

Predicted 3D models are improved through a refinement process that focuses on correcting irregularities in their structure, including unusual bonds, torsion angles, and hydrogen bonding patterns, to align them better with the native structure. Different types of restraints have been effectively used in molecular dynamics (MD) simulation-based refinement approaches since CASP10. CASP 12 included the evaluation of ReFOLD, a molecular dynamics-based refinement approach by the McGuffin group, alongside other comparable methods.