Unlocking the antiviral arsenal: Computational exploration of Langya virus (L, C, V, W) through phylogenetic analysis and molecular modeling

INTRODUCTION

The emergence of the Langya virus (LayV) poses a significant public health threat, particularly in Eastern China, with the potential for widespread transmission and pandemic outbreaks. Initially detected in individuals with animal exposure and fever, LayV has led to 100 confirmed cases between 2018 and 2023.^[1] With no approved treatments available, it is imperative to unravel the virus’s mechanisms and develop effective countermeasures. Four crucial LayV proteins – L, C, V, and W – play pivotal roles in viral infection and dissemination. The L protein drives virus RNA synthesis, while the C protein regulates replication and assembly. In addition, the V protein aids in particle production and release, and the W protein facilitates immune evasion.^[2,3] A thorough investigation into the origins of LayV indicates a likelihood of animal transmission, with shrews being identified as potential carriers. Yet, additional research is necessary to confirm this hypothesis. At present, there is no evidence of human-to-human transmission, underscoring the importance of cautious hygiene practices, especially following animal contact.^[4]

LayV infections initially mimic flu-like symptoms, including fever, fatigue, coughing, and muscle pain, with rare instances of severe manifestations. LayV is related to other dangerous viruses such as Hernipavirus and Nipah, but it does not seem as deadly right now, yet continuous monitoring and research are important. Scientists are leveraging LayV proteins for diagnostics and therapeutics, developing targeted tests focusing on specific protein domains, and exploring antibody-based surveillance methods. Scientists are studying a new virus called LayV. They are looking closely at the proteins of this virus to develop ways to diagnose and treat infections.^[5] For diagnosis, they are working on tests that can quickly and accurately detect LayV infection by focusing on specific parts of the virus proteins. They are also studying antibodies against these proteins to better understand how the virus spreads and affects people. In terms of treatment, they are looking at potential drugs that can target the activities of two important LayV proteins, L and W. However, figuring out the details of these proteins is tricky, so designing effective drugs is a complex task. Therapeutically, efforts are directed toward drugs targeting the L protein to inhibit virus replication and the W protein to mitigate immune interference. However, the intricacies of LayV protein interactions necessitate meticulous drug design to combat potential resistance. Understanding the pressing need for an antiviral against the LayV, this research initiative employs a strategic approach. Initially, it delves into the virus’s evolutionary history to grasp its origins and transmission patterns. Next, it utilizes advanced tools such as the Swiss Model (SM)^[6] and AlphaFold (AF)^[7] to examine a crucial virus protein in three dimensions. This detailed analysis of the protein’s structure is essential for understanding the viruses mechanisms and for informed therapeutic development.

MATERIALS AND METHODS

Active site prediction for modeled protein: Leveraging advanced computational tools

In the pursuit of elucidating crucial insights into the modeled protein, a strategic approach to active site prediction was employed, tapping into cutting-edge computational resources. The state-of-the-art Computed Atlas of Surface Topography of Proteins (CASTp) 3.0,^[21] PHYRE2 Protein Fold Recognition and Analysis Server,^[22] PrankWeb,^[23] and SCFBio^[24] played pivotal roles in this endeavor. CASTp, a dynamic online server, was harnessed to identify and quantify voids within the intricate 3D structure of the protein. Simultaneously, PHYRE2, a robust online web server recognized for its prowess in predicting and analyzing protein structure, function, and mutations, was utilized.

The protein’s 3D model was seamlessly uploaded to the CASTp and PHYRE2. We pinpointed the spatial locations of potential active sites and garnered insights into the essential amino acids orchestrating binding interactions.^[25] This approach, merging computational efficiency with sophisticated algorithms, adds a layer of precision to our exploration of the protein’s functional domains, laying the groundwork for a nuanced understanding of its biological significance [Figure 1].

Close

Figure 1:

Comprehensive workflow for protein analysis and modeling. MEGA: Molecular evolutionary genetics analysis, PROCHE: Protein contact heterogeneity, PROSA: Protein structure analysis, CASTp: Computed atlas of surface topography of proteins, PHYRE2: Protein homology/analogy recognition engine V 2.0, SCF: Stem cell factor.

Export to PPT

RESULTS

The nucleotide sequences associated with the 16S rDNA region of the isolated forms have been meticulously archived in Table 1 of the NCBI database. Noteworthy is the analysis of LayV sp strains within the scope of this investigation, revealing a sequence homology spanning from 96% to 99% with LayV SDQD_S1801, as documented in the NCBI database. This assessment was conducted employing stringent criteria, including the consideration of the lowest E-value, maximum query coverage, and highest identity. Models characterized by the lowest Bayesian information criterion (BIC) scores were selected to best elucidate the substitution pattern. Within this framework, the K2 model, representing Kimura’s 2 parameters, demonstrated the most favorable BIC scores, registering at 228080.2 [Table 1A]. The consensus ML tree, depicted in Figure 2, was derived from the alignment of these sequences.

Close

Figure 2:

Illustrates the Maximum Likelihood relationship among various species in the Worldwide collection, with a focus on Langya Virus sp (LayV). LayV is intricately associated with paramyxovirus, mojinga virus, hernipavirus, and melian virus within this context. MAG: Metagenome-assembled genome, GHA: Genome-Wide haplotypic association, SYS: Systems biology, RNA: Ribonucleic acid.

Export to PPT

The phylogenetic analysis was conducted employing MEGA5 software, wherein a multiple alignment file served as the foundational dataset. The analysis, encompassing 70 nucleotide sequences, aimed at elucidating the evolutionary relationships among the strains. The Kimura 2-parameter model, integrated with the ML method, facilitated the inference of the evolutionary history. Robustness of the results was ensured through a bootstrap analysis of 1000 replicates. To initiate the heuristic search, automatic generation of initial trees involved the application of the Neighbour-Joining and BioNJ algorithms were applied to a pairwise distance matrix generated using the Maximum Composite Likelihood method. The selection of the topology with the highest log-likelihood value further refined the analysis. Notably, positions featuring gaps or missing data were meticulously eliminated to enhance data reliability. The resultant phylogenetic tree, presented in Figure 2, exhibited a discernible clustering pattern. Specifically, all native strains formed an evolutionary cluster alongside Paramyxovirus, while other LayV species were distinctly grouped based on relatedness. The tree, drawn to scale, expressed branch lengths in substitutions per site, providing a comprehensive visual representation of the evolutionary landscape.

The mean (relative) evolutionary rates, calculated using the Kimura 2-parameter model, were normalized to the average evolutionary rate across all sites. Consequently, sites with a rate <1 are evolving at a pace slower than the average, while those with a rate >1 are evolving more rapidly than the overall average. The estimated value of the shape parameter for the discrete gamma distribution is 0.5674. Substitution pattern and rates were estimated under the Tamura-Nei (1993) model (+G). A discrete gamma distribution was used to model evolutionary rate differences among sites (5 categories, [+G]). Mean evolutionary rates in these categories were 0.03, 0.19, 0.52, 1.14, 3.12 substitutions per site. The nucleotide frequencies are A = 34.66%, T/U = 28.45%, C = 17.19%, and G = 19.70%. For estimating ML values, a tree topology was automatically computed. The maximum log likelihood for this computation was −114066.116. This analysis involved 11 nucleotide sequences. Codon positions included were 1^st+2^nd+3^rd+noncoding [Figure 3]. There were a total of 21134 positions in the final dataset. Evolutionary analyses were conducted in MEGA X.

Close

Figure 3:

Representation of nucleotide frequencies (A, T, G, C) for constructing Maximum Likelihood (ML) phylogeny. A,T,G,C stands for adenine (A), thymine (T), guanine (G), and cytosine (C).

Export to PPT

The results obtained from Tajima’s test for a set of three sequences reveal valuable insights into the genetic variation among them. Among the scrutinized sites, a substantial number of 9718 were identified as identical across all three sequences, indicating a high level of conservation [Table 2]. Conversely, 177 sites exhibited divergence in all three sequences, suggesting instances of genetic variation within the shared regions. Notably, Sequence A displayed 6523 unique differences, implying a distinct genetic makeup, while Sequence B and Sequence C presented 166 and 203 unique differences, respectively. These unique differences underscore the individualistic genetic signatures of each sequence. This comprehensive analysis provides a nuanced understanding of both shared and distinctive genetic features among the three sequences, contributing to a more holistic interpretation of their genetic relationships.

The molecular evolution analysis of the Paramyxovirus sp population demonstrated a close relationship with LayV, providing robust support for the native strain’s topology. Within this investigation, the concept of selection pressure driving molecular evolution was posited. The findings of the current analysis highlighted a noteworthy trend: A significant proportion of sites in the 16S rDNA region evolved at a markedly slower rate than the average. In addition, the nucleotide diversity across the entire population was observed to be very low, and the Tajima test statistic value (D) showed a slight positive inclination. These results suggest two plausible scenarios: Either the population may have undergone a recent bottleneck or is experiencing a decrease, or there might be evidence pointing to over-dominant selection at this specific locus. These nuanced insights contribute to a more comprehensive understanding of the dynamic forces shaping the molecular evolution of the paramyxovirus sp population.

Homology modeling

In the pursuit of unraveling the intricacies of LayV proteins (C, L, V, W), structures were generated from the corresponding FASTA sequences, derived both from the target proteins and template structures chosen for homology modeling, a rigorous examination ensued using SWISS-MODEL and AF. This encompassing analysis delved into various quality metrics, including the precision of folding, identification of steric clashes among unpaired atoms, and the residue-wise stereochemical integrity of the protein structure, as depicted in Figure 4. The subsequent phase involved a detailed scrutiny of the structural data, initially focusing on the accuracy of predicting functional sites, such as binding sites. This step was crucial in assessing the model’s reliability in capturing biologically relevant features. Subsequently, the efficacy of both in silico approaches was meticulously evaluated, considering their overall performance in generating reliable protein structures.

Close

Figure 4:

Displays three-dimensional modeled structures of (a) C, (b) L, (c) V, (d) W proteins of Langya Virus generated using Swiss Model. In this representation, α-helices, β-strands, and loops are distinctly colored in red, blue, and green, respectively.

Export to PPT

Comparative analysis of homology models and AF structures: Assessing structural integrity and discrepancies

The validation of the model, encompassing the geometric properties of the backbone conformations, was meticulously scrutinized through diverse structure evaluation programs. Figure 5 illustrates the Ramachandran plot for the three models, with the SM showcasing notable results. This model evaluates the permissible torsional angles for amino acid residues in protein backbones. Specifically, C and L exhibit the most favorable outcomes, with 93.2% and 89.4% of residues residing in favored regions, respectively. In contrast, V and W also display best result, with 82.0% and 87.5%, respectively. This insightful analysis enhances our understanding of the model’s structural integrity and informs potential refinements for optimal performance.

Close

Figure 5:

Illustrates the Ramachandran Plot post-refinement for the modeled proteins, including (a) C protein, (b) L protein, (c) V protein, and (d) W protein of Langya Virus. In this depiction, the favored, allowed, and disallowed regions are represented by red, yellow, and white regions, respectively. This visualization provides a comprehensive overview of the torsional angles of amino acid residues in the refined protein structures.

Export to PPT

Table 3 provides a thorough comparative analysis of four protein models (C, L, V, and W) employing the ERRAT and Verify_3D validation tools, offering valuable insights into packing quality and adherence to physical-chemical principles. Notably, C and L emerge as frontrunners with impressive ERRAT scores of 87.5% and 89.2% [Figure 6], indicating superior packing and efficient residue interactions. The corresponding Verify_3D scores of 86.5% and 85.4% further substantiate their robust structural integrity. V presents a nuanced scenario with a commendable ERRAT score of 88.46 but a slightly lower Verify_3D score of 87.3, suggesting potential packing or chemical inconsistencies. Conversely, W raises concerns with significantly lower ERRAT (77.4) and Verify_3D (78.4) scores, underscoring stability issues. In summary, C and L showcase high structural integrity, V demands additional scrutiny, and W warrants further investigation and potential refinement for enhanced accuracy. The inclusion of ProSA [Figures 7-9] and QE assessments reinforces the SM’s superiority compared to other proteins, underscoring its efficacy in structural development [Figure 10].

Table 3: A multitool validation landscape for all protein models.

Name of the Protein	Validations		After modeling	Refine loop	Minimize	Predict side chain
C Protein	Ramachandran Plot	Favored regions	93.2	93.2	90.2	88.4
		Additional allowed regions	5.1	3	2	3
		Generously allowed regions	0	0	0	0
		Disallowed regions	1.7	1	1	1.2
	Errat		87.5	91	95.1	89.1
	Verify_3D		86.5	88.5	88.1	85
	Prove_z-score		0.6	0.5	0.58	0.54
L Protein	Ramachandran Plot	Favored regions	89.4	90.2	88.5	87.5
		Additional allowed regions	7.1	6.8	6.5	6.9
		Generously allowed regions	2.4	1.7	2.5	2.6
		Disallowed regions	1.2	1.1	1.7	1.8
	Errat		89.2	89.9	88.5	87.4
	Verify_3D		85.4	84.8	84.5	86.7
	Prove_z-score		0.7	0.73	0.68	0.74
V Protein	Ramachandran Plot	Favored regions	82.0	83	82.5	81.8
		Additional allowed regions	12.0	11	11.5	12.5
		Generously allowed regions	4.0	3	2	4.4
		Disallowed regions	2.0	3	4	2.1
	Errat		88.46	88.32	87.5	85.9
	Verify_3D		87.3	86.9	87.4	85.5
	Prove_z-score		0.66	0.67	0.59	0.62
W Protein	Ramachandran Plot	Favored regions	87.5	87.9	87.2	86.8
		Additional allowed regions	12.5	12.1	12.8	12.8
		Generously allowed regions	0.0	0.1	0	2.4
		Disallowed regions	0.0	0	0	0.1
	Errat		77.4	78.4	77.2	77.5
	Verify_3D		78.4	78.9	77.5	77.3
	Prove_z-score		0.422	0.45	0.54	0.45

This table presents a detailed analysis of your models’ structural integrity, assessed through a battery of validation tools. Procheck, ProQ, and ERRAT zoom in on backbone angles, overall quality, and packing efficiency, respectively. Verify_3D ensures adherence to physical principles, while Prove and ProSA scrutinize “protein-likeness” and potential errors. Finally, Z-scores benchmark your models against known structures. Comparing scores across these tools reveals each model’s strengths and weaknesses, paving the way for informed refinement and a deeper understanding of their structural validity

Close

Figure 6:

ERRAT plots paint a stark contrast: (a) C and (b) L bask in the high-confidence zones, their packing near-flawless, while (c) V dips into uncertainty, and (d) W plunges, suggesting potential instability. This packing quality spectrum highlights the need for tailored validation across protein models. *: Likely indicates that the “Error value” represents a specific measure of structural quality, such as deviation from ideal geometry or model accuracy. The exact metric is typically defined in the methods or legend. White bars: These may represent residues within an acceptable quality threshold, indicating areas of the protein structure with lower error values and, thus, reliable structural quality. Yellow bars: Likely highlight residues with higher error values, suggesting regions in the protein structure with potential issues or instability, which may need further refinement or indicate areas of flexibility.

Export to PPT

Close

Figure 7:

Displays residue energy plots for proteins generated by the Swiss model, with specific representations for (a) C, (b) L, (c) V, and (d) W proteins. The green lines indicate residue energy levels.

Export to PPT

Close

Figure 8:

ProSA-web Z-scores and Frequency for the Swiss model (indicated by black spots) concerning (a) C, (b) L, (c) V, and (d) W proteins, (e) C protein, (f) L protein, (g) V protein and (h) W protein. These Z-scores are compared to all protein chains in the protein data bank determined through X-ray crystallography (represented in light blue) or NMR spectroscopy (depicted in dark blue), considering their respective lengths.

Export to PPT

Close

Figure 9:

Verify_3D plots for proteins generated by the Swiss model, with specific representations for (a) C, (b) L, (c) V, and (d) W proteins.

Export to PPT

Close

Figure 10:

The analysis encompasses side chain prediction accompanied by quality estimate plots for proteins generated through the Swiss model. This investigation provides detailed insights into residue quality, with dedicated representations for (a) C, (b) L, (c) V, and (d) W proteins. Orange Bars: These represent regions of lower predicted local similarity to the target structure. Blue Bars: These represent regions of Higher predicted local similarity to the target structure.

Export to PPT

DISCUSSION

Evaluation of the AlphaFold structural models

The C protein

The structural analysis reveals the outstanding quality of the C protein model generated by Swiss, boasting an impressive 91% accuracy in the allowed region and demonstrating fewer disallowed regions compared to the AlphaFold structure [Figure 11]. This exceptional performance is substantiated through a detailed examination, particularly emphasized in Figure 12 Ramachandran plot. The plot vividly illustrates the C protein model’s adherence to preferred torsional angles for amino acid residues, showcasing a predominant presence in the allowed region, indicative of its robust structural integrity. In contrast, the AlphaFold structure exhibits a lower accuracy in the allowed region, emphasizing the superior conformational accuracy achieved by the SM. This comprehensive analysis reinforces the conclusion that the C protein model from Swiss excels in both accuracy and minimized disallowed regions, establishing its superiority over the AlphaFold structure in terms of structural quality.

Close

Figure 11:

Displays three-dimensional modeled structures of (a) C, (b) L, (c) V, and (d) W proteins of Langya Virus generated using Alpha Fold. The representation highlights α-helices in red, β-strands in yellow, and loops in blue and green, providing a visually distinct and informative depiction of the protein structures.

Export to PPT

Close

Figure 12:

The Ramachandran plot depicts the post-refinement status of the modeled proteins, encompassing (a) C protein, (b) L protein, (c) V Protein, and (d) W protein. Within this representation, the favored, allowed, and disallowed regions are denoted by red, yellow, and white areas, respectively. This visualization offers a comprehensive overview of the torsional angles of amino acid residues within the refined protein structures.

Export to PPT

The L protein

In the structural analysis, the L protein model crafted by AF’s structure stands out as the optimal choice, boasting an impressive 89.5% accuracy in the allowed region and exhibiting fewer disallowed regions. Furthermore, the model’s predicted local distance difference test (pLDDT) score of 81.5 serves as a robust confidence measure, indicating the model’s predicted accuracy on the local Distance Difference Test for Cα (Alpha Carbon) (lDDT-Xα) metric [Table 4]. This score is instrumental in assessing the reliability of different regions within the model. Notably, regions with pLDDT >90 are deemed highly accurate and suitable for applications requiring precision, such as characterizing binding sites. Regions with pLDDT between 70 and 90 are anticipated to have a generally good backbone prediction, while those between 50 and 70 warrant caution due to lower confidence. Regions with pLDDT <50, often exhibiting a ribbon-like appearance, are indicative of potential disorder and should be approached with care. This nuanced understanding, as detailed in different analyses, including the Ramachandran plot, affirms the robustness and reliability of the L protein model produced by AF’s structure.

Table 4: Comparative values of Procheck, Hydrophobic, and pLDDT analysis of Template and Modeled proteins of all the four models.

Name of the Protein	Validations		After modeling
C Protein	Ramachandran Plot	Favored regions	71.2
		Additional allowed regions	19.4
		Generously allowed regions	7.5
		Disallowed regions	1.9
	Hydrophobic Plot		2.3
	pLDDT		47.8
L Protein	Ramachandran Plot	Favored regions	89.5
		Additional allowed regions	7.1
		Generously allowed regions	2.4
		Disallowed regions	1.2
	Hydrophobic plot		2.5
	pLDDT		81.5
V Protein	Ramachandran plot	Favored regions	51.8
		Additional allowed regions	37.8
		Generously allowed regions	7.7
		Disallowed regions	2.8
	Hydrophobic Plot		3.8
	pLDDT		44.5
W Protein	Ramachandran Plot	Favored regions	52.8
		Additional allowed regions	36.8
		Generously allowed regions	7.7
		Disallowed regions	2.8
	Hydrophobic Plot		3.5
	pLDDT		41.5

pLDDT: Predicted Local Distance Difference Test

The V protein

The V protein model crafted by the SM emerges as the optimal choice, showcasing an impressive 82.5% accuracy in the allowed region and demonstrating fewer disallowed regions. In addition, the model’s pLDDT score of 44.5 provides a confidence measure, though relatively lower, emphasizing its predicted accuracy on the lDDT-Cα metric. Notably, a comprehensive evaluation of Figures 13-16 suggests that the AF analysis figure yields lower residue estimation quality compared to the residue plot estimation of the V protein model produced by the SM, as depicted in Figure 6. This comparative assessment underscores the robustness and reliability of the V protein model generated by the SM across various analyses. As a result, we have chosen this particular model for the continuation of further studies, particularly in the realm of docking studies.

Close

Figure 13:

Presents a hydrophobic plot with distinct representations for (a) C, (b) L, (c) V, and (d) W proteins, all modeled using AlphaFold within Discovery Studio.

Export to PPT

Close

Figure 14:

Sequence identity plot with distinct representations for (a) C, (b) L, (c) V, and (d) W proteins, all modeled using AlphaFold.

Export to PPT

Close

Figure 15:

Displays residue energy plots for proteins generated by the AlphaFold, with specific representations for (a) C, (b) L, (c) V, and (d) W proteins.

Export to PPT

Close

Figure 16:

Represents a contact plot of side chain estimation residue with distinct representations for (a) C, (b) L, (c) V, and (d) W proteins, all modeled using AlphaFold within Discovery Studio.

Export to PPT

The W protein

Comparison of the W protein model generated by SM structure revealed superior performance in several key metrics. The model achieved an impressive 87.5% accuracy in predicting allowed regions, exhibiting a notably lower rate of disallowed regions compared to alternative approaches. Notably, the pLDDT score of 43.5 signifies enhanced model confidence. Visual analysis of Figures 13 and 14 further corroborates these findings, demonstrating reduced hydrophobic activity (3.5) and lower sequence similarity, both suggestive of a more accurate and refined model.

A thorough comparative analysis of protein structures generated by SM and AF reveals distinct strengths inherent to each platform. SM excels in the modeling of Proteins C, V, and W, showcasing higher accuracy and more precise template selection when compared to AF [Figures 15 and 16]. This proficiency can be attributed to the robust algorithm of SM, particularly well suited for homology modeling based on sequence similarity. Conversely, AF proves highly effective for Protein L, demonstrating precise residue count and favorable Z-score and root mean square deviation values, as detailed in Table 5. Recognizing these complementary strengths, the choice of an optimal structure for further studies in antiviral drug development should be approached on a case-by-case basis, taking into consideration the specific protein in question and the desired level of accuracy and detail required for the investigation.

Table 5: Comparative values of Z-score mean, RMSD, and Z-score standard deviation in different stages of refinement used in Swiss model and AlphaFold software.

Name of software	Name of protein	Z−score mean	RMSD	Z−score standard deviation
Swiss model	C Protein	0.521	1.531	1.445
	L Protein	0.65	1.42	1.3
	V Protein	0.67	1.46	1.35
	W Protein	0.4	1.34	1.28
AlphaFold	C Protein	0.75	1.34	1.2
	L Protein	0.83	1.1	1.45
	V Protein	0.78	1	1.57
	W Protein	0.82	0.89	1.79

RMSD: Root mean square deviation

Active site prediction

In a recent study, researchers employed structural bioinformatics to unravel the potential binding sites of four fascinating proteins: C, L, V, and W. By leveraging the CASTp, PrankWeb, and SCFBio server for structural comparisons and binding site predictions, alongside template-based modeling and integrating experimental binding data, the scientists shed light on key amino acid residues crucial for substrate interaction [Figures 17-19].

Close

Figure 17:

Active site prediction of the target proteins C, L, V, and W of the Langya virus using CASTp (Computed Atlas of Surface Topography of Proteins).

Export to PPT

Close

Figure 18:

Predicted ligand binding sites for target proteins C, L, V, and W of Langya virus using PrankWeb.

Export to PPT

Close

Figure 19:

Active sites for target proteins C, L, V, and W of langay virus obtained from the Supercomputing Facility for Bioinformatics and Computational Biology, with visualization using Discovery Studio.

Export to PPT

For Protein C, conserved residues LEU 105, ASN 109, and ILE 125 emerged as favorable docking sites, aligning perfectly with existing data on CD1 and OD1 binding. Similar insights were gleaned for Protein L, where a pocket area of 5920.328 Å2 and residues CA 38, CB 40, and OG 40 emerged as potential binding partners, echoing the interaction profile with HIS and SER. In Protein V, a conserved pocket area and residues NH2, CB, and CD were pinpointed as promising docking sites, corroborating findings on ARG, GLU, and TYR involvement. Finally, for Protein W, the computational analysis unraveled TYR, ARG, and PRO residues as potential docking partners [Figure 17].

This comprehensive understanding of binding interactions in these proteins holds immense significance. It lays the groundwork for rational drug design, paves the way for elucidating functional mechanisms, and opens doors for novel therapeutic strategies. The future lies in experimentally validating these predictions through site-directed mutagenesis and structural biology techniques, while simultaneously exploring the functional implications of these binding interactions in diverse biological contexts. This research not only illuminates the power of structural bioinformatics in identifying potential binding sites but also sparks further experimentation and advances our knowledge of protein function, paving the way for exciting discoveries in the future.

CONCLUSION

This research study focused on the computational exploration of LayV through phylogenetic analysis and molecular modeling. By integrating advanced methodologies such as phylogenetic analysis and molecular modeling, the study aimed to unravel the biological intricacies of LayV and accelerate the development of antiviral therapeutics. The phylogenetic analysis provided insights into the evolutionary trajectory of LayV, revealing its close kinship with other henipaviruses such as the Wenzhou shrew henipavirus 1 and Mojiang virus. The analysis also highlighted the genetic diversity of LayV and its potential implications for future research and public health. Through homology modeling, the study predicted the 3D structures of LayV proteins L, C, V, and W. The models were validated using rigorous evaluation criteria, including structural analysis, energy analysis, and compatibility analysis. The results identified the SwissModel-generated homology models for proteins C, V, and W as superior, with the V Model AF standing out as optimal. Furthermore, the study employed advanced computational tools for active site prediction, allowing for the identification and quantification of active sites within the protein structures. This information is crucial for understanding the functional domains and potential binding interactions of LayV proteins.

Overall, this research provides a comprehensive computational approach to unlocking the antiviral arsenal against LayV. The findings pave the way for targeted antiviral interventions and contribute to the development of diagnostic tools and therapeutic interventions. By elucidating the molecular choreography of LayV and its proteins, this study reinforces the commitment to public health and addresses the complexities of viral infections.