A machine learning method for the identification and characterization of novel COVID-19 drug targets.

In addition to vaccines, the World Health Organization sees novel medications as an urgent matter to fight the ongoing COVID-19 pandemic. One possible strategy is to identify target proteins, for which a perturbation by an existing compound is likely to benefit COVID-19 patients. In order to contribute to this effort, we present GuiltyTargets-COVID-19 ( https://guiltytargets-covid.eu/ ), a machine learning supported web tool to identify novel candidate drug targets. Using six bulk and three single cell RNA-Seq datasets, together with a lung tissue specific protein-protein interaction network, we demonstrate that GuiltyTargets-COVID-19 is capable of (i) prioritizing meaningful target candidates and assessing their druggability, (ii) unraveling their linkage to known disease mechanisms, (iii) mapping ligands from the ChEMBL database to the identified targets, and (iv) pointing out potential side effects in the case that the mapped ligands correspond to approved drugs. Our example analyses identified 4 potential drug targets from the datasets: AKT3 from both the bulk and single cell RNA-Seq data as well as AKT2, MLKL, and MAPK11 in the single cell experiments. Altogether, we believe that our web tool will facilitate future target identification and drug development for COVID-19, notably in a cell type and tissue specific manner.

De novo design of protein interactions with learned surface fingerprints.

Physical interactions between proteins are essential for most biological processes governing life1. However, the molecular determinants of such interactions have been challenging to understand, even as genomic, proteomic and structural data increase. This knowledge gap has been a major obstacle for the comprehensive understanding of cellular protein-protein interaction networks and for the de novo design of protein binders that are crucial for synthetic biology and translational applications2-9. Here we use a geometric deep-learning framework operating on protein surfaces that generates fingerprints to describe geometric and chemical features that are critical to drive protein-protein interactions10. We hypothesized that these fingerprints capture the key aspects of molecular recognition that represent a new paradigm in the computational design of novel protein interactions. As a proof of principle, we computationally designed several de novo protein binders to engage four protein targets: SARS-CoV-2 spike, PD-1, PD-L1 and CTLA-4. Several designs were experimentally optimized, whereas others were generated purely in silico, reaching nanomolar affinity with structural and mutational characterization showing highly accurate predictions. Overall, our surface-centric approach captures the physical and chemical determinants of molecular recognition, enabling an approach for the de novo design of protein interactions and, more broadly, of artificial proteins with function.

Probing Ligand Binding Sites on Large Proteins by Nuclear Magnetic Resonance Spectroscopy of Genetically Encoded Non-Canonical Amino Acids.

N6-(((trimethylsilyl)-methoxy)carbonyl)-l-lysine (TMSK) and N6-trifluoroacetyl-l-lysine (TFAK) are non-canonical amino acids, which can be installed in proteins by genetic encoding. In addition, we describe a new aminoacyl-tRNA synthetase specific for N6-(((trimethylsilyl)methyl)-carbamoyl)-l-lysine (TMSNK), which is chemically more stable than TMSK. Using the dimeric SARS-CoV-2 main protease (Mpro) as a model system with three different ligands, we show that the 1H and 19F nuclei of the solvent-exposed trimethylsilyl and CF3 groups produce intense signals in the nuclear magnetic resonance (NMR) spectrum. Their response to active-site ligands differed significantly when positioned near rather than far from the active site. Conversely, the NMR probes failed to confirm the previously reported binding site of the ligand pelitinib, which was found to enhance the activity of Mpro by promoting the formation of the enzymatically active dimer. In summary, the amino acids TMSK, TMSNK, and TFAK open an attractive path for site-specific NMR analysis of ligand binding to large proteins of limited stability and at low concentrations.

Metabolomic and Proteomic Profiling of Porcine Intestinal Epithelial Cells Infected with Porcine Epidemic Diarrhea Virus.

Porcine epidemic diarrhea virus (PEDV) infection results in severe epidemic diarrhea and the death of suckling pigs. Although new knowledge about the pathogenesis of PEDV has been improved, alterations in metabolic processes and the functional regulators involved in PEDV infection with host cells remain largely unknow. To identify cellular metabolites and proteins related to PEDV pathogenesis, we synergistically investigated the metabolome and proteome profiles of PEDV-infected porcine intestinal epithelial cells by liquid chromatography tandem mass spectrometry and isobaric tags for relative and absolute quantification techniques. We identified 522 differential metabolites in positive and negative ion modes and 295 differentially expressed proteins after PEDV infection. Pathways of cysteine and methionine metabolism, glycine, serine and threonine metabolism, and mineral absorption were significantly enriched by differential metabolites and differentially expressed proteins. The betaine-homocysteine S-methyltransferase (BHMT) was indicated as a potential regulator involved in these metabolic processes. We then knocked down the BHMT gene and observed that down-expression of BHMT obviously decreased copy numbers of PEDV and virus titers (p < 0.01). Our findings provide new insights into the metabolic and proteomic profiles in PEDV-infected host cells and contribute to our further understanding of PEDV pathogenesis.

AI-accelerated protein-ligand docking for SARS-CoV-2 is 100-fold faster with no significant change in detection.

Protein-ligand docking is a computational method for identifying drug leads. The method is capable of narrowing a vast library of compounds down to a tractable size for downstream simulation or experimental testing and is widely used in drug discovery. While there has been progress in accelerating scoring of compounds with artificial intelligence, few works have bridged these successes back to the virtual screening community in terms of utility and forward-looking development. We demonstrate the power of high-speed ML models by scoring 1 billion molecules in under a day (50 k predictions per GPU seconds). We showcase a workflow for docking utilizing surrogate AI-based models as a pre-filter to a standard docking workflow. Our workflow is ten times faster at screening a library of compounds than the standard technique, with an error rate less than 0.01% of detecting the underlying best scoring 0.1% of compounds. Our analysis of the speedup explains that another order of magnitude speedup must come from model accuracy rather than computing speed. In order to drive another order of magnitude of acceleration, we share a benchmark dataset consisting of 200 million 3D complex structures and 2D structure scores across a consistent set of 13 million "in-stock" molecules over 15 receptors, or binding sites, across the SARS-CoV-2 proteome. We believe this is strong evidence for the community to begin focusing on improving the accuracy of surrogate models to improve the ability to screen massive compound libraries 100 × or even 1000 × faster than current techniques and reduce missing top hits. The technique outlined aims to be a fast drop-in replacement for docking for screening billion-scale molecular libraries.

ZDHHC11 Suppresses Zika Virus Infections by Palmitoylating the Envelope Protein.

Zika virus (ZIKV) is an RNA-enveloped virus that belongs to the Flavivirus genus, and ZIKV infections potentially induce severe neurodegenerative diseases and impair male fertility. Palmitoylation is an important post-translational modification of proteins that is mediated by a series of DHHC-palmitoyl transferases, which are implicated in various biological processes and viral infections. However, it remains to be investigated whether palmitoylation regulates ZIKV infections. In this study, we initially observed that the inhibition of palmitoylation by 2-bromopalmitate (2-BP) enhanced ZIKV infections, and determined that the envelope protein of ZIKV is palmitoylated at Cys308. ZDHHC11 was identified as the predominant enzyme that interacts with the ZIKV envelope protein and catalyzes its palmitoylation. Notably, ZDHHC11 suppressed ZIKV infections in an enzymatic activity-dependent manner and ZDHHC11 knockdown promoted ZIKV infection. In conclusion, we proposed that the envelope protein of ZIKV undergoes a novel post-translational modification and identified a distinct mechanism in which ZDHHC11 suppresses ZIKV infections via palmitoylation of the ZIKV envelope protein.

DeepGpgs: a novel deep learning framework for predicting arginine methylation sites combined with Gaussian prior and gated self-attention mechanism.

Protein arginine methylation is an important posttranslational modification (PTM) associated with protein functional diversity and pathological conditions including cancer. Identification of methylation binding sites facilitates a better understanding of the molecular function of proteins. Recent developments in the field of deep neural networks have led to a proliferation of deep learning-based methylation identification studies because of their fast and accurate prediction. In this paper, we propose DeepGpgs, an advanced deep learning model incorporating Gaussian prior and gated attention mechanism. We introduce a residual network channel to extract the evolutionary information of proteins. Then we combine the adaptive embedding with bidirectional long short-term memory networks to form a context-shared encoder layer. A gated multi-head attention mechanism is followed to obtain the global information about the sequence. A Gaussian prior is injected into the sequence to assist in predicting PTMs. We also propose a weighted joint loss function to alleviate the false negative problem. We empirically show that DeepGpgs improves Matthews correlation coefficient by 6.3% on the arginine methylation independent test set compared with the existing state-of-the-art methylation site prediction methods. Furthermore, DeepGpgs has good robustness in phosphorylation site prediction of SARS-CoV-2, which indicates that DeepGpgs has good transferability and the potential to be extended to other modification sites prediction. The open-source code and data of the DeepGpgs can be obtained from https://github.com/saizhou1/DeepGpgs.

Isthmin-A Multifaceted Protein Family.

Isthmin (ISM) is a secreted protein family with two members, namely ISM1 and ISM2, both containing a TSR1 domain followed by an AMOP domain. Its broad expression pattern suggests diverse functions in developmental and physiological processes. Over the past few years, multiple studies have focused on the functional analysis of the ISM protein family in several events, including angiogenesis, metabolism, organ homeostasis, immunity, craniofacial development, and cancer. Even though ISM was identified two decades ago, we are still short of understanding the roles of the ISM protein family in embryonic development and other pathological processes. To address the role of ISM, functional studies have begun but unresolved issues remain. To elucidate the regulatory mechanism of ISM, it is crucial to determine its interactions with other ligands and receptors that lead to the activation of downstream signalling pathways. This review provides a perspective on the gene organization and evolution of the ISM family, their links with developmental and physiological functions, and key questions for the future.

Myeloperoxidase-derived hypochlorous acid targets human airway epithelial plasmalogens liberating protein modifying electrophilic 2-chlorofatty aldehydes.

Neutrophil and airway epithelial cell interactions are critical in the inflammatory response to viral infections including respiratory syncytial virus, Sendai virus, and SARS-CoV-2. Airway epithelial cell dysfunction during viral infections is likely mediated by the interaction of virus and recruited neutrophils at the airway epithelial barrier. Neutrophils are key early responders to viral infection. Neutrophil myeloperoxidase catalyzes the conversion of hydrogen peroxide to hypochlorous acid (HOCl). Previous studies have shown HOCl targets host neutrophil and endothelial cell plasmalogen lipids, resulting in the production of the chlorinated lipid, 2-chlorofatty aldehyde (2-ClFALD). We have previously shown that the oxidation product of 2-ClFALD, 2-chlorofatty acid (2-ClFA) is present in bronchoalveolar lavage fluid of Sendai virus-infected mice, which likely results from the attack of the epithelial plasmalogen by neutrophil-derived HOCl. Herein, we demonstrate small airway epithelial cells contain plasmalogens enriched with oleic acid at the sn-2 position unlike endothelial cells which contain arachidonic acid enrichment at the sn-2 position of plasmalogen. We also show neutrophil-derived HOCl targets epithelial cell plasmalogens to produce 2-ClFALD. Further, proteomics and over-representation analysis using the ω-alkyne analog of the 2-ClFALD molecular species, 2-chlorohexadecanal (2-ClHDyA) showed cell adhesion molecule binding and cell-cell junction enriched categories similar to that observed previously in endothelial cells. However, in contrast to endothelial cells, proteins in distinct metabolic pathways were enriched with 2-ClFALD modification, particularly pyruvate metabolism was enriched in epithelial cells and mitochondrial pyruvate respiration was reduced. Collectively, these studies demonstrate, for the first time, a novel plasmalogen molecular species distribution in airway epithelial cells that are targeted by myeloperoxidase-derived hypochlorous acid resulting in electrophilic 2-ClFALD, which potentially modifies epithelial physiology by modifying proteins.

Rapid Reversible Osmoregulation of Cytoplasmic Biomolecular Condensates of Human Interferon-α-Induced Antiviral MxA GTPase.

We previously discovered that exogenously expressed GFP-tagged cytoplasmic human myxovirus resistance protein (MxA), a major antiviral effector of Type I and III interferons (IFNs) against several RNA- and DNA-containing viruses, existed in the cytoplasm in phase-separated membraneless biomolecular condensates of varying sizes and shapes with osmotically regulated disassembly and reassembly. In this study we investigated whether cytoplasmic IFN-α-induced endogenous human MxA structures were also biomolecular condensates, displayed hypotonic osmoregulation and the mechanisms involved. Both IFN-α-induced endogenous MxA and exogenously expressed GFP-MxA formed cytoplasmic condensates in A549 lung and Huh7 hepatoma cells which rapidly disassembled within 1-2 min when cells were exposed to 1,6-hexanediol or to hypotonic buffer (~40-50 mOsm). Both reassembled into new structures within 1-2 min of shifting cells to isotonic culture medium (~330 mOsm). Strikingly, MxA condensates in cells continuously exposed to culture medium of moderate hypotonicity (in the range one-fourth, one-third or one-half isotonicity; range 90-175 mOsm) first rapidly disassembled within 1-3 min, and then, in most cells, spontaneously reassembled 7-15 min later into new structures. This spontaneous reassembly was inhibited by 2-deoxyglucose (thus, was ATP-dependent) and by dynasore (thus, required membrane internalization). Indeed, condensate reassembly was preceded by crowding of the cytosolic space by large vacuole-like dilations (VLDs) derived from internalized plasma membrane. Remarkably, the antiviral activity of GFP-MxA against vesicular stomatitis virus survived hypoosmolar disassembly and subsequent reassembly. The data highlight the exquisite osmosensitivity of MxA condensates, and the preservation of antiviral activity in the face of hypotonic stress.

Evolutionary Signals in Coronaviral Structural Proteins Suggest Possible Complex Mechanisms of Post-Translational Regulation in SARS-CoV-2 Virus.

Post-translational regulation of proteins has emerged as a central topic of research in the field of functional proteomics. Post-translational modifications (PTMs) dynamically control the activities of proteins and are involved in a wide range of biological processes. Crosstalk between different types of PTMs represents a key mechanism of regulation and signaling. Due to the current pandemic of the novel and dangerous SARS-CoV-2 (Severe Acute Respiratory Syndrome Coronavirus 2) virus, here we present an in silico analysis of different types of PTMs in structural proteins of coronaviruses. A dataset of PTM sites was studied at three levels: conservation analysis, mutational analysis and crosstalk analysis. We identified two sets of PTMs which could have important functional roles in the regulation of the structural proteins of coronaviruses. Additionally, we found seven interesting signals of potential crosstalk events. These results reveal a higher level of complexity in the mechanisms of post-translational regulation of coronaviral proteins and provide new insights into the adaptation process of the SARS-CoV-2 virus.

Activation mechanisms of monocytes/macrophages in adult-onset Still disease.

Adult onset Still disease (AOSD) is a systemic inflammatory disorder characterized by skin rash, spiking fever, arthritis, sore throat, lymphadenopathy, and hepatosplenomegaly. Although the etiology of this disease has not been fully clarified, both innate and acquired immune responses could contribute to its pathogenesis. Hyperactivation of macrophages and neutrophils along with low activation of natural killer (NK) cells in innate immunity, as well as hyperactivation of Th1 and Th17 cells, whereas low activation of regulatory T cells (Tregs) in acquired immunity are involved in the pathogenic process of AOSD. In innate immunity, activation of monocytes/macrophages might play central roles in the development of AOSD and macrophage activation syndrome (MAS), a severe life-threating complication of AOSD. Regarding the activation mechanisms of monocytes/macrophages in AOSD, in addition to type II interferon (IFN) stimulation, several pathways have recently been identified, such as the pathogen-associated molecular patterns (PAMPs) and damage-associated molecular patterns (DAMPs)-pattern recognition receptors (PRRs) axis, and neutrophil extracellular traps (NETs)-DNA. These stimulations on monocytes/macrophages cause activation of the nucleotide-binding oligomerization domain, leucine-rich repeat, and pyrin domain (NLRP) 3 inflammasomes, which trigger capase-1 activation, resulting in conversion of pro-IL-1ß and pro-IL-18 into mature forms. Thereafter, IL-1ß and IL-18 produced by activated monocytes/macrophages contribute to various clinical features in AOSD. We identified placenta-specific 8 (PLAC8) as a specifically increased molecule in monocytes of active AOSD, which correlated with serum levels of CRP, ferritin, IL-1ß, and IL-18. Interestingly, PLAC8 could suppress the synthesis of pro-IL-1ß and pro-IL-18 via enhanced autophagy; thus, PLAC8 seems to be a regulatory molecule in AOSD. These findings for the activation mechanisms of monocytes/macrophages could shed light on the pathogenesis and development of a novel therapeutic strategy for AOSD.

Lysosomal enzyme trafficking factor LYSET enables nutritional usage of extracellular proteins.

Mammalian cells can generate amino acids through macropinocytosis and lysosomal breakdown of extracellular proteins, which is exploited by cancer cells to grow in nutrient-poor tumors. Through genetic screens in defined nutrient conditions, we characterized LYSET, a transmembrane protein (TMEM251) selectively required when cells consume extracellular proteins. LYSET was found to associate in the Golgi with GlcNAc-1-phosphotransferase, which targets catabolic enzymes to lysosomes through mannose-6-phosphate modification. Without LYSET, GlcNAc-1-phosphotransferase was unstable because of a hydrophilic transmembrane domain. Consequently, LYSET-deficient cells were depleted of lysosomal enzymes and impaired in turnover of macropinocytic and autophagic cargoes. Thus, LYSET represents a core component of the lysosomal enzyme trafficking pathway, underlies the pathomechanism for hereditary lysosomal storage disorders, and may represent a target to suppress metabolic adaptations in cancer.

Rapidly evolving viral motifs mostly target biophysically constrained binding pockets of host proteins.

Evolutionary changes in host-virus interactions can alter the course of infection, but the biophysical and regulatory constraints that shape interface evolution remain largely unexplored. Here, we focus on viral mimicry of host-like motifs that allow binding to host domains and modulation of cellular pathways. We observe that motifs from unrelated viruses preferentially target conserved, widely expressed, and highly connected host proteins, enriched with regulatory and essential functions. The interface residues within these host domains are more conserved and bind a larger number of cellular proteins than similar motif-binding domains that are not known to interact with viruses. In contrast, rapidly evolving viral-binding human proteins form few interactions with other cellular proteins and display high tissue specificity, and their interfaces have few inter-residue contacts. Our results distinguish between conserved and rapidly evolving host-virus interfaces and show how various factors limit host capacity to evolve, allowing for efficient viral subversion of host machineries.

Thiol Modifications in the Extracellular Space-Key Proteins in Inflammation and Viral Infection.

Posttranslational modifications (PTMs) allow to control molecular and cellular functions in response to specific signals and changes in the microenvironment of cells. They regulate structure, localization, stability, and function of proteins in a spatial and temporal manner. Among them, specific thiol modifications of cysteine (Cys) residues facilitate rapid signal transduction. In fact, Cys is unique because it contains the highly reactive thiol group that can undergo different reversible and irreversible modifications. Upon inflammation and changes in the cellular microenvironment, many extracellular soluble and membrane proteins undergo thiol modifications, particularly dithiol-disulfide exchange, S-glutathionylation, and S-nitrosylation. Among others, these thiol switches are essential for inflammatory signaling, regulation of gene expression, cytokine release, immunoglobulin function and isoform variation, and antigen presentation. Interestingly, also the redox state of bacterial and viral proteins depends on host cell-mediated redox reactions that are critical for invasion and infection. Here, we highlight mechanistic thiol switches in inflammatory pathways and infections including cholera, diphtheria, hepatitis, human immunodeficiency virus (HIV), influenza, and coronavirus disease 2019 (COVID-19).

PHILM2Web: A high-throughput database of macromolecular host-pathogen interactions on the Web.

During infection, the pathogen's entry into the host organism, breaching the host immune defense, spread and multiplication are frequently mediated by multiple interactions between the host and pathogen proteins. Systematic studying of host-pathogen interactions (HPIs) is a challenging task for both experimental and computational approaches and is critically dependent on the previously obtained knowledge about these interactions found in the biomedical literature. While several HPI databases exist that manually filter HPI protein-protein interactions from the generic databases and curated experimental interactomic studies, no comprehensive database on HPIs obtained from the biomedical literature is currently available. Here, we introduce a high-throughput literature-mining platform for extracting HPI data that includes the most comprehensive to date collection of HPIs obtained from the PubMed abstracts. Our HPI data portal, PHILM2Web (Pathogen-Host Interactions by Literature Mining on the Web), integrates an automatically generated database of interactions extracted by PHILM, our high-precision HPI literature-mining algorithm. Currently, the database contains 23 581 generic HPIs between 157 host and 403 pathogen organisms from 11 609 abstracts. The interactions were obtained from processing 608 972 PubMed abstracts, each containing mentions of at least one host and one pathogen organisms. In response to the coronavirus disease 2019 (COVID-19) pandemic, we also utilized PHILM to process 25 796 PubMed abstracts obtained by the same query as the COVID-19 Open Research Dataset. This COVID-19 processing batch resulted in 257 HPIs between 19 host and 31 pathogen organisms from 167 abstracts. The access to the entire HPI dataset is available via a searchable PHILM2Web interface; scientists can also download the entire database in bulk for offline processing. Database URL: http://philm2web.live.

Comparing protein-protein interaction networks of SARS-CoV-2 and (H1N1) influenza using topological features.

SARS-CoV-2 pandemic first emerged in late 2019 in China. It has since infected more than 298 million individuals and caused over 5 million deaths globally. The identification of essential proteins in a protein-protein interaction network (PPIN) is not only crucial in understanding the process of cellular life but also useful in drug discovery. There are many centrality measures to detect influential nodes in complex networks. Since SARS-CoV-2 and (H1N1) influenza PPINs pose 553 common human proteins. Analyzing influential proteins and comparing these networks together can be an effective step in helping biologists for drug-target prediction. We used 21 centrality measures on SARS-CoV-2 and (H1N1) influenza PPINs to identify essential proteins. We applied principal component analysis and unsupervised machine learning methods to reveal the most informative measures. Appealingly, some measures had a high level of contribution in comparison to others in both PPINs, namely Decay, Residual closeness, Markov, Degree, closeness (Latora), Barycenter, Closeness (Freeman), and Lin centralities. We also investigated some graph theory-based properties like the power law, exponential distribution, and robustness. Both PPINs tended to properties of scale-free networks that expose their nature of heterogeneity. Dimensionality reduction and unsupervised learning methods were so effective to uncover appropriate centrality measures.

A close shave: How SARS-CoV-2 induces the loss of cilia.

Wang et al. report in this issue (2022. J. Cell Biol.https://doi.org/10.1083/jcb.202108015) that the SARS-CoV-2 protein ORF10 increases the activity of the E3 ligase CUL2ZYG11B, leading to the degradation of multiple ciliary proteins. The resulting loss of cilia may facilitate the spread of SARS-CoV-2 in the respiratory tree.

Targeted protein degradation: from small molecules to complex organelles-a Keystone Symposia report.

Targeted protein degradation is critical for proper cellular function and development. Protein degradation pathways, such as the ubiquitin proteasomes system, autophagy, and endosome-lysosome pathway, must be tightly regulated to ensure proper elimination of misfolded and aggregated proteins and regulate changing protein levels during cellular differentiation, while ensuring that normal proteins remain unscathed. Protein degradation pathways have also garnered interest as a means to selectively eliminate target proteins that may be difficult to inhibit via other mechanisms. On June 7 and 8, 2021, several experts in protein degradation pathways met virtually for the Keystone eSymposium "Targeting protein degradation: from small molecules to complex organelles." The event brought together researchers working in different protein degradation pathways in an effort to begin to develop a holistic, integrated vision of protein degradation that incorporates all the major pathways to understand how changes in them can lead to disease pathology and, alternatively, how they can be leveraged for novel therapeutics.

Network for network concept offers new insights into host- SARS-CoV-2 protein interactions and potential novel targets for developing antiviral drugs.

SARS-CoV-2, the causal agent of COVID-19, is primarily a pulmonary virus that can directly or indirectly infect several organs. Despite many studies carried out during the current COVID-19 pandemic, some pathological features of SARS-CoV-2 have remained unclear. It has been recently attempted to address the current knowledge gaps on the viral pathogenicity and pathological mechanisms via cellular-level tropism of SARS-CoV-2 using human proteomics, visualization of virus-host protein-protein interactions (PPIs), and enrichment analysis of experimental results. The synergistic use of models and methods that rely on graph theory has enabled the visualization and analysis of the molecular context of virus/host PPIs. We review current knowledge on the SARS-COV-2/host interactome cascade involved in the viral pathogenicity through the graph theory concept and highlight the hub proteins in the intra-viral network that create a subnet with a small number of host central proteins, leading to cell disintegration and infectivity. Then we discuss the putative principle of the "gene-for-gene and "network for network" concepts as platforms for future directions toward designing efficient anti-viral therapies.