In Silico Analysis of Spike Protein Mutations and Their Impact on Host Receptor Affinity: A Computational Virology Approach

Introduction

The spike glycoprotein of enveloped viruses mediates host cell entry by binding to specific cellular receptors and facilitating membrane fusion. In coronaviruses, the receptor-binding domain (RBD) within the S1 subunit is the primary determinant of host tropism and zoonotic potential [1, 2, 3]. Mutations in the RBD can alter binding affinity to host receptors such as angiotensin-converting enzyme 2 (ACE2), thereby modulating infectivity, cross-species transmission, and immune evasion [4, 5, 34]. Computational virology provides a powerful framework to predict the structural and biophysical consequences of such mutations before experimental validation, enabling rapid risk assessment for emerging variants [6, 7, 8].

This article presents an exhaustive review of in silico methodologies applied to spike protein mutation analysis, with emphasis on receptor-binding affinity predictions, molecular dynamics simulations, docking studies, and machine learning-based classification. While the primary model system discussed is SARS-CoV-2, the principles extend directly to veterinary coronaviruses such as feline coronavirus (FCoV), bovine coronavirus (BCoV), and bat-derived sarbecoviruses, for which similar computational pipelines are employed [9, 10, 11]. The goal is to provide a technical reference for veterinary virologists and computational biologists engaged in zoonotic risk assessment and antiviral design.

Computational Methods for Spike Protein Mutation Analysis

Molecular Dynamics Simulations

Molecular dynamics (MD) simulations model the time-dependent behavior of spike protein variants at atomic resolution. By applying force fields such as CHARMM or AMBER, MD trajectories reveal conformational changes in the RBD induced by point mutations [12, 13, 14]. Key metrics extracted from MD include root-mean-square deviation (RMSD), root-mean-square fluctuation (RMSF), and radius of gyration (Rg), which collectively indicate structural stability and flexibility [15, 16]. For example, simulations of the D614G mutation demonstrated increased RBD opening and enhanced ACE2 accessibility, correlating with higher infectivity [17, 18]. Free energy landscapes derived from principal component analysis (PCA) further identify metastable states that favor receptor engagement [19, 14].

Molecular Docking and Binding Free Energy Calculations

Docking algorithms predict the preferred orientation of a ligand (e.g., RBD) within a receptor binding site. Rigid and flexible docking protocols, implemented in tools such as AutoDock Vina and Rosetta, generate binding poses and estimate binding affinities through scoring functions [10, 11, 20]. More accurate predictions are obtained using end-point free energy methods like Molecular Mechanics Poisson-Boltzmann Surface Area (MM-PBSA) and Molecular Mechanics Generalized Born Surface Area (MM-GBSA), which decompose binding energy into van der Waals, electrostatic, and solvation contributions [10, 12, 13]. These calculations have been applied to quantify the impact of N501Y, E484K, and K417N mutations on ACE2 affinity [21, 22, 23].

Machine Learning and Deep Learning Approaches

Machine learning models trained on large mutational datasets can predict variant properties directly from sequence or structure. Contrastive learning frameworks, such as those described by Tenekeci et al., efficiently classify viral escape mutations by comparing mutational neighborhoods [1]. Structure-based predictors using graph neural networks or random forests on residue contact maps have achieved high accuracy in forecasting binding affinity changes [2, 3]. Deep mutational scanning data, combined with supervised learning, enables the identification of synergistic mutation combinations that confer immune evasion [21, 24]. These approaches are increasingly integrated into surveillance pipelines for early detection of high-risk variants [25, 26].

Immunoinformatics and Epitope Prediction

Immunoinformatic tools predict T-cell and B-cell epitopes within the spike protein and assess how mutations alter recognition by host immune receptors. NetMHCpan and similar algorithms evaluate peptide-MHC binding affinities, while structural modeling of antibody-RBD interfaces identifies escape mutations that reduce neutralization [7, 25, 27]. Coevolution-based methods further account for epitope variability across strains, guiding the design of broadly protective vaccines [24]. For veterinary applications, these tools are adapted to species-specific MHC haplotypes and antibody repertoires [26, 28].

Key Spike Protein Mutations and Their Functional Consequences

D614G

The D614G substitution in the S1/S2 junction region became dominant early in the SARS-CoV-2 pandemic. MD simulations revealed that this mutation stabilizes the open conformation of the RBD, increasing ACE2 binding probability [17, 18]. Docking studies confirmed a modest but significant enhancement in binding affinity, correlating with higher viral loads in animal models [3, 5]. This mutation also alters the furin cleavage site dynamics, potentially affecting cell entry efficiency [22].

N501Y

The N501Y mutation, located at the RBD-ACE2 interface, introduces a tyrosine residue that forms additional pi-pi stacking interactions with Y41 of ACE2. MM-PBSA calculations showed a 2- to 3-fold increase in binding free energy compared to the wild-type RBD [13, 21]. This mutation is a hallmark of the Alpha and Omicron lineages and is associated with expanded host tropism, including enhanced binding to murine and canine ACE2 orthologs [4, 34].

E484K

E484K reduces electrostatic complementarity at the RBD-ACE2 interface, leading to decreased binding affinity in some contexts. However, its primary impact is on antibody escape, as the residue lies within a major neutralizing epitope [21, 22]. Computational alanine scanning and docking of monoclonal antibodies demonstrated that E484K disrupts hydrogen bonding and salt bridges, reducing neutralization potency by over 10-fold [7, 11, 20].

K417N/T

Mutations at position 417 (K417N in Beta and K417T in Delta) alter the electrostatic potential of the RBD surface. K417N reduces binding affinity to ACE2 by approximately 30% in MM-GBSA calculations, but this loss is often compensated by other mutations in the same variant [13, 23]. The mutation also affects recognition by class 1 neutralizing antibodies, contributing to immune escape [25, 21].

Omicron-Specific Mutations

The Omicron variant (B.1.1.529) carries over 30 mutations in the spike protein, including G339D, S371L, S373P, S375F, K417N, N440K, G446S, S477N, T478K, E484A, Q493R, G496S, Q498R, N501Y, and Y505H. Comprehensive MD and docking studies revealed that these mutations collectively enhance ACE2 binding while simultaneously reducing antibody recognition [13, 4, 16]. The Q498R and N501Y pair, for example, creates a new salt bridge with ACE2 residue D38, increasing affinity beyond that of the ancestral strain [21, 23]. Recombination events among Omicron subvariants further diversify the spike protein, complicating receptor binding predictions [4].

Impact on Host Receptor Affinity and Zoonotic Potential

ACE2 Binding Dynamics

ACE2 is the primary receptor for SARS-CoV-2 and related sarbecoviruses. In silico analyses have systematically evaluated the binding affinity of spike RBD variants to ACE2 orthologs from various animal species, including bats, pangolins, cats, dogs, ferrets, and cattle [29, 5, 34]. These studies identify key residues in ACE2 that determine cross-species compatibility. For instance, residue H34 in human ACE2 is critical for binding, whereas the corresponding residue in feline ACE2 (N34) reduces affinity for some variants [29, 34]. Such predictions inform risk assessments for reverse zoonosis and reservoir establishment.

Alternative Receptors and Co-receptors

Beyond ACE2, spike proteins can engage alternative receptors such as DC-SIGN, DC-SIGNR, and neuropilin-1 (NRP1). Gupta et al. used in silico docking to show that Omicron spike variants exploit DC-SIGN/DC-SIGNR for enhanced attachment, potentially contributing to immune evasion and severity [30]. Similarly, NRP1 binding is influenced by mutations in the furin cleavage site, as predicted by molecular docking of single nucleotide polymorphisms in the NRP1 gene [35]. These interactions expand the host cell tropism and may facilitate spillover into species with low ACE2 affinity.

Predicting Spillover Risk

Integrative computational pipelines combine sequence surveillance, structural modeling, and binding affinity predictions to assess zoonotic risk. For example, Park et al. developed a comprehensive model that accurately predicted infectivity of SARS-CoV-2 variants based on spike sequence features [3]. Machine learning classifiers trained on deep mutational scanning data can identify mutations that simultaneously enhance ACE2 binding and reduce antibody neutralization, marking them as high-risk candidates [1, 2]. These frameworks are directly transferable to animal coronaviruses, such as those circulating in bats and pangolins, to forecast spillover events [4, 5].

Tools and Databases for In Silico Spike Analysis

Structural Prediction and Modeling

AlphaFold2 and Rosetta are widely used for predicting spike protein structures and modeling mutant conformations. AlphaFold2 provides high-accuracy models of the full-length spike trimer, while Rosetta enables flexible backbone refinement and side-chain repacking upon mutation [11, 31, 15]. These tools have been applied to design stabilized spike proteins for vaccine development and to engineer nanobodies with enhanced affinity [7, 20, 31].

Sequence Databases and Surveillance Platforms

GISAID (Global Initiative on Sharing All Influenza Data) serves as the primary repository for SARS-CoV-2 genomic sequences, enabling real-time tracking of spike mutations [32, 33]. The Protein Data Bank (PDB) and UniProt provide experimentally determined structures and functional annotations for spike proteins and receptors [34]. Computational pipelines routinely fetch sequences from GISAID, map mutations to PDB structures, and perform batch docking or MD simulations [3, 16].

Workflow Integration

A typical in silico analysis workflow for spike protein mutations is depicted in Figure 1.

flowchart TD
    A[Sequence Surveillance GISAID], > B[Mutation Identification]
    B, > C[Structural Modeling AlphaFold2 / Rosetta]
    C, > D[Molecular Dynamics Simulations]
    C, > E[Molecular Docking]
    D, > F[Binding Free Energy MM-PBSA / MM-GBSA]
    E, > F
    F, > G[Affinity Prediction]
    B, > H[Machine Learning Classifier]
    H, > G
    G, > I[Risk Assessment Zoonotic / Immune Escape]
    I, > J[Experimental Validation]

Figure 1. Integrated computational workflow for predicting the impact of spike protein mutations on host receptor affinity and zoonotic risk.

Tables of Key Mutations and Predicted Effects

Table 1. Common Spike Mutations and Their Predicted Impact on ACE2 Binding Affinity

Table 2. Computational Methods and Their Applications in Spike Mutation Analysis

Conclusion

In silico analysis of spike protein mutations has become an indispensable component of veterinary virology and zoonotic risk assessment. By integrating molecular dynamics, docking, free energy calculations, and machine learning, researchers can rapidly evaluate the functional consequences of emerging mutations on host receptor affinity and immune escape. The methodologies reviewed here, validated extensively on SARS-CoV-2, are directly applicable to animal coronaviruses and other zoonotic pathogens. Continued development of computational tools and databases will further enhance our ability to predict spillover events and guide intervention strategies.

References

[1] Tenekeci S, Sezgin E, Tekir S. A Contrastive Learning Framework for Efficient Viral Escape Prediction. IEEE Trans Comput Biol Bioinform. 2026. https://pubmed.ncbi.nlm.nih.gov/41843532/

[2] van den Boom M, Schultes E, Hankemeier T. Structure-based prediction of SARS-CoV-2 variant properties using machine learning on mutational neighborhoods. Front Bioinform. 2025. https://pubmed.ncbi.nlm.nih.gov/40989750/

[3] Park J, Choi W, Seong DY, et al. Accurate predictions of SARS-CoV-2 infectivity from comprehensive analysis. Elife. 2024. https://pubmed.ncbi.nlm.nih.gov/39717902/

[4] Arbi M, Khedhiri M, Ayouni K, et al. Recombination Events Among SARS-CoV-2 Omicron Subvariants: Impact on Spike Interaction With ACE2 Receptor and Neutralizing Antibodies. Evol Bioinform Online. 2024. https://pubmed.ncbi.nlm.nih.gov/39149136/

[5] Zaheer M, Ali N, Javed H, et al. Uncovering the impact of SARS-CoV2 spike protein variants on human receptors: A molecular dynamics docking and simulation approach. J Infect Public Health. 2023. https://pubmed.ncbi.nlm.nih.gov/37566991/

[6] Hensel RC, Materón EM, Macedo AJG, et al. Nanoyeast-based impedimetric biosensor with mutated single chain antigen-binding fragment anchoring for SARS-CoV-2 detection. Biomed Microdevices. 2026. https://pubmed.ncbi.nlm.nih.gov/41770397/

[7] Yeo J, Yun MR, Kim SY, et al. Structure-guided design of a bivalent SARS-CoV-2 mRNA vaccine with NTD stabilizing mutations enhances broad immunity. Front Immunol. 2025. https://pubmed.ncbi.nlm.nih.gov/41660617/

[8] Hurmach V, Karaushu V, Prylutska S, et al. In Silico Analysis of C(60) Fullerene Interaction with TMPRSS2: Toward Novel COVID-19 Prevention Approaches. Molecules. 2025. https://pubmed.ncbi.nlm.nih.gov/41375182/

[9] Ahmad R, Akhter Y, Varsha, et al. In Silico and in vitro evaluation of Cucurbitaceae seed lectins as promising prophylactic and therapeutic agents against SARS-CoV-2. Glycoconj J. 2025. https://pubmed.ncbi.nlm.nih.gov/41205137/

[10] Ahuja P, Rao L, Dsouza N. Targeting SARS-CoV-2 variants with Bromelain: an in-silico study of Spike-ACE2 inhibition using docking, molecular dynamic simulation, and MM-PBSA. In Silico Pharmacol. 2025. https://pubmed.ncbi.nlm.nih.gov/41113175/

[11] Cao S, Sun B, Gao F. From immune evasion to broad in silico binding: computational optimization of SARS-CoV-2 RBD-targeting nanobody. Front Immunol. 2025. https://pubmed.ncbi.nlm.nih.gov/40895570/

[12] Kalhor H, Mokhtarian MH, Rahimi H, et al. Predictive Insights Into Bioactive Compounds from Streptomyces as Inhibitors of SARS-CoV-2 Mutant Strains by Receptor Binding Domain: Molecular Docking and Dynamics Simulation Approaches. Iran J Pharm Res. 2024. https://pubmed.ncbi.nlm.nih.gov/40066112/

[13] Haider S, Ahmad N, Shafiq M, et al. Uncovering the Binding Mechanism of Mutated Omicron Variants via Computational Strategies. ACS Omega. 2025. https://pubmed.ncbi.nlm.nih.gov/39895721/

[14] Elseginy SA. Exploring binding mechanisms of omicron spike protein with dolutegravir and etravirine by molecular dynamics simulation, principal component analysis, and free binding energy calculations. J Biomol Struct Dyn. 2025. https://pubmed.ncbi.nlm.nih.gov/38084778/

[15] Unlu S, Uskudar-Guclu A, Cela I. The impacts of 13 novel mutations of SARS-CoV-2 on protein dynamics: In silico analysis from Turkey. Hum Gene (Amst). 2022. https://pubmed.ncbi.nlm.nih.gov/37520163/

[16] Pritam M, Dutta S, Medicherla KM, et al. Computational analysis of spike protein of SARS-CoV-2 (Omicron variant) for development of peptide-based therapeutics and diagnostics. J Biomol Struct Dyn. 2024. https://pubmed.ncbi.nlm.nih.gov/37498146/

[17] Patel P, Kaushik N, Acharya TR, et al. Inactivation of Pseudovirus Expressing the D614G Spike Protein Mutation using Nitric Oxide-Plasma Activated Water. Adv Sci (Weinh). 2024. https://pubmed.ncbi.nlm.nih.gov/39535372/

[18] Zargan S, Jalili H, Dabirmanesh B, et al. Amyloidogenesis of SARS-CoV-2 delta plus and omicron variants receptor-binding domain (RBD): impact of SUMO fusion tag. Biotechnol Lett. 2024. https://pubmed.ncbi.nlm.nih.gov/39182215/

[19] Prasad R, Kadam A, Padippurackal VV, et al. Discovery of small molecule entry inhibitors targeting the linoleic acid binding pocket of SARS-CoV-2 spike protein. J Biomol Struct Dyn. 2025. https://pubmed.ncbi.nlm.nih.gov/38520147/

[20] Singh V, Bhutkar M, Choudhary S, et al. Structure-guided mutations in CDRs for enhancing the affinity of neutralizing SARS-CoV-2 nanobody. Biochem Biophys Res Commun. 2024. https://pubmed.ncbi.nlm.nih.gov/39366179/

[21] Pham LV, Underwood AP, Binderup A, et al. Neutralisation resistance of SARS-CoV-2 spike-variants is primarily mediated by synergistic receptor binding domain substitutions. Emerg Microbes Infect. 2024. https://pubmed.ncbi.nlm.nih.gov/39392057/

[22] El-Assaad AM, Hamieh T. SARS-CoV-2: Prediction of critical ionic amino acid mutations. Comput Biol Med. 2024. https://pubmed.ncbi.nlm.nih.gov/38870723/

[23] Ovchynnykova O, Booth JD, Cocroft TM, et al. In silico Study on Natural Chemical Compounds from Citric Essential Oils as Potential Inhibitors of an Omicron (BA.1) SARS-CoV-2 Mutants' Spike Glycoprotein. Curr Comput Aided Drug Des. 2025. https://pubmed.ncbi.nlm.nih.gov/38178668/

[24] Mishra SK, Priya P, Rai GP, et al. Coevolution based immunoinformatics approach considering variability of epitopes to combat different strains: A case study using spike protein of SARS-CoV-2. Comput Biol Med. 2023. https://pubmed.ncbi.nlm.nih.gov/37422941/

[25] da Silva GCS, Paraná VC, de Almeida Rego FF, et al. Impact of mutations in immunodominant regions of SARS-CoV-2 variants on recognition by CD8+ T cell: An in silico analysis. J Infect Public Health. 2025. https://pubmed.ncbi.nlm.nih.gov/40359819/

[26] Kumar KM, Karthik Y, Ramakrishna D, et al. Immunoinformatic exploration of a multi-epitope-based peptide vaccine candidate targeting emerging variants of SARS-CoV-2. Front Microbiol. 2023. https://pubmed.ncbi.nlm.nih.gov/37915849/

[27] Vecchio E, Rotundo S, Veneziano C, et al. The spike-specific TCRβ repertoire shows distinct features in unvaccinated or vaccinated patients with SARS-CoV-2 infection. J Transl Med. 2024. https://pubmed.ncbi.nlm.nih.gov/38185632/

[28] Subhadra B, Agrawal R, Pal VK, et al. Significant Broad-Spectrum Antiviral Activity of Bi121 against Different Variants of SARS-CoV-2. Viruses. 2023. https://pubmed.ncbi.nlm.nih.gov/37376598/

[29] Shafique R, Zahra A, Butt MH, et al. Comparative analysis of ACE2 mutations and their impact on SARS-CoV-2 susceptibility within diverse ethnic groups. J Biomol Struct Dyn. 2026. https://pubmed.ncbi.nlm.nih.gov/40063087/

[30] Gupta J, Malik MZ, Chaturvedi M, et al. SARS CoV-2 spike protein variants exploit DC-SIGN/DC-SIGNR receptor for evolution

[31] Shang W, Hu X, Lin X, et al. Iterative In Silico Screening for Optimizing Stable Conformation of Anti-SARS-CoV-2 Nanobodies. Pharmaceuticals (Basel). 2024. https://pubmed.ncbi.nlm.nih.gov/38675386/

[32] Olivero NB, Zappia VE, Gargantini P, et al. Genomic Evolution of the SARS-CoV-2 Omicron Variant in Córdoba, Argentina (2021-2022): Analysis of Uncommon and Prevalent Spike Mutations. Viruses. 2024. https://pubmed.ncbi.nlm.nih.gov/39772187/

[33] Gopalan V, Chandran A, Arumugam K, et al. Distribution and Functional Analyses of Mutations in Spike Protein and Phylogenic Diversity of SARS-CoV-2 Variants Emerged during the Year 2021 in India. J Glob Infect Dis. 2023. https://pubmed.ncbi.nlm.nih.gov/37469462/