Extracting prime protein targets as possible drug candidates: machine learning evaluation.

Extracting "high ranking" or "prime protein targets" (PPTs) as potent MRSA drug candidates from a given set of ligands is a key challenge in efficient molecular docking. This study combines protein-versus-ligand matching molecular docking (MD) data extracted from 10 independent molecular docking (MD) evaluations - ADFR, DOCK, Gemdock, Ledock, Plants, Psovina, Quickvina2, smina, vina, and vinaxb to identify top MRSA drug candidates. Twenty-nine active protein targets (APT) from the enhanced DUD-E repository ( http://DUD-E.decoys.org ) are matched against 1040 ligands using "forward modeling" machine learning for initial "data mining and modeling" (DDM) to extract PPTs and the corresponding high affinity ligands (HALs). K-means clustering (KMC) is then performed on 400 ligands matched against 29 PTs, with each cluster accommodating HALs, and the corresponding PPTs. Performance of KMC is then validated against randomly chosen head, tail, and middle active ligands (ALs). KMC outcomes have been validated against two other clustering methods, namely, Gaussian mixture model (GMM) and density based spatial clustering of applications with noise (DBSCAN). While GMM shows similar results as with KMC, DBSCAN has failed to yield more than one cluster and handle the noise (outliers), thus affirming the choice of KMC or GMM. Databases obtained from ADFR to mine PPTs are then ranked according to the number of the corresponding HAL-PPT combinations (HPC) inside the derived clusters, an approach called "reverse modeling" (RM). From the set of 29 PTs studied, RM predicts high fidelity of 5 PPTs (17%) that bind with 76 out of 400, i.e., 19% ligands leading to a prediction of next-generation MRSA drug candidates: PPT2 (average HPC is 41.1%) is the top choice, followed by PPT14 (average HPC 25.46%), and then PPT15 (average HPC 23.12%). This algorithm can be generically implemented irrespective of pathogenic forms and is particularly effective for sparse data.

Towards Effective Consensus Scoring in Structure-Based Virtual Screening.

Virtual screening (VS) is a computational strategy that uses in silico automated protein docking inter alia to rank potential ligands, or by extension rank protein-ligand pairs, identifying potential drug candidates. Most docking methods use preferred sets of physicochemical descriptors (PCDs) to model the interactions between host and guest molecules. Thus, conventional VS is often data-specific, method-dependent and with demonstrably differing utility in identifying candidate drugs. This study proposes four universality classes of novel consensus scoring (CS) algorithms that combine docking scores, derived from ten docking programs (ADFR, DOCK, Gemdock, Ledock, PLANTS, PSOVina, QuickVina2, Smina, Autodock Vina and VinaXB), using decoys from the DUD-E repository ( http://dude.docking.org/ ) against 29 MRSA-oriented targets to create a general VS formulation that can identify active ligands for any suitable protein target. Our results demonstrate that CS provides improved ligand-protein docking fidelity when compared to individual docking platforms. This approach requires only a small number of docking combinations and can serve as a viable and parsimonious alternative to more computationally expensive docking approaches. Predictions from our CS algorithm are compared against independent machine learning evaluations using the same docking data, complementing the CS outcomes. Our method is a reliable approach for identifying protein targets and high-affinity ligands that can be tested as high-probability candidates for drug repositioning.

To Affinity and Beyond: A Personal Reflection on the Design and Discovery of Drugs.

Faced with new and as yet unmet medical need, the stark underperformance of the pharmaceutical discovery process is well described if not perfectly understood. Driven primarily by profit rather than societal need, the search for new pharmaceutical products-small molecule drugs, biologicals, and vaccines-is neither properly funded nor sufficiently systematic. Many innovative approaches remain significantly underused and severely underappreciated, while dominant methodologies are replete with problems and limitations. Design is a component of drug discovery that is much discussed but seldom realised. In and of itself, technical innovation alone is unlikely to fulfil all the possibilities of drug discovery if the necessary underlying infrastructure remains unaltered. A fundamental revision in attitudes, with greater reliance on design powered by computational approaches, as well as a move away from the commercial imperative, is thus essential to capitalise fully on the potential of pharmaceutical intervention in healthcare.

West Nile Virus Vaccine Design by T Cell Epitope Selection: In Silico Analysis of Conservation, Functional Cross-Reactivity with the Human Genome, and Population Coverage.

West Nile Virus (WNV) causes a debilitating and life-threatening neurological disease in humans. Since its emergence in Africa 50 years ago, new strains of WNV and an expanding geographical distribution have increased public health concerns. There are no licensed therapeutics against WNV, limiting effective infection control. Vaccines represent the most efficacious and efficient medical intervention known. Epitope-based vaccines against WNV remain significantly underexploited. Here, we use a selection protocol to identify a set of conserved prevalidated immunogenic T cell epitopes comprising a putative WNV vaccine. Experimentally validated immunogenic WNV epitopes and WNV sequences were retrieved from the IEDB and West Nile Virus Variation Database. Clustering and multiple sequence alignment identified a smaller subset of representative sequences. Protein variability analysis identified evolutionarily conserved sequences, which were used to select a diverse set of immunogenic candidate T cell epitopes. Cross-reactivity and human leukocyte antigen-binding affinities were assessed to eliminate unsuitable epitope candidates. Population protection coverage (PPC) quantified individual epitopes and epitope combinations against the world population. 3 CD8+ T cell epitopes (ITYTDVLRY, TLARGFPFV, and SYHDRRWCF) and 1 CD4+ epitope (VTVNPFVSVATANAKVLI) were selected as a putative WNV vaccine, with an estimated PPC of 97.14%.

Correction to: Computational assembly of a human Cytomegalovirus vaccine upon experimental epitope legacy.

After publication of the original article [1], we were notified that legends of Fig. 1 and Fig. 2 have been swapped.

Drug Discovery: Today and Tomorrow.

Drug discovery continues to underperform relative to unmet medical need. Driven by profit not societal need, the search for new drugs is neither properly funded nor sufficiently systematic. Many innovative approaches are significantly underused yet extant methodology is replete with problems. In and of itself, technical innovation is unlikely to fulfill the potential of drug discovery if the supporting infrastructure remains unchanged.

Computational assembly of a human Cytomegalovirus vaccine upon experimental epitope legacy.

BACKGROUND: Human Cytomegalovirus (HCMV) is a ubiquitous herpesvirus affecting approximately 90% of the world population. HCMV causes disease in immunologically naive and immunosuppressed patients. The prevention, diagnosis and therapy of HCMV infection are thus crucial to public health. The availability of effective prophylactic and therapeutic treatments remain a significant challenge and no vaccine is currently available. Here, we sought to define an epitope-based vaccine against HCMV, eliciting B and T cell responses, from experimentally defined HCMV-specific epitopes. RESULTS: We selected 398 and 790 experimentally validated HCMV-specific B and T cell epitopes, respectively, from available epitope resources and apply a knowledge-based approach in combination with immunoinformatic predictions to ensemble a universal vaccine against HCMV. The T cell component consists of 6 CD8 and 6 CD4 T cell epitopes that are conserved among HCMV strains. All CD8 T cell epitopes were reported to induce cytotoxic activity, are derived from early expressed genes and are predicted to provide population protection coverage over 97%. The CD4 T cell epitopes are derived from HCMV structural proteins and provide a population protection coverage over 92%. The B cell component consists of just 3 B cell epitopes from the ectodomain of glycoproteins L and H that are highly flexible and exposed to the solvent. CONCLUSIONS: We have defined a multiantigenic epitope vaccine ensemble against the HCMV that should elicit T and B cell responses in the entire population. Importantly, although we arrived to this epitope ensemble with the help of computational predictions, the actual epitopes are not predicted but are known to be immunogenic.

VaxiJen Dataset of Bacterial Immunogens: An Update.

BACKGROUND: Identifying immunogenic proteins is the first stage in vaccine design and development. VaxiJen is the most widely used and highly cited server for immunogenicity prediction. As the developers of VaxiJen, we are obliged to update and improve it regularly. Here, we present an updated dataset of bacterial immunogens containing 317 experimentally proven immunogenic proteins of bacterial origin, of which 60% have been reported during the last 10 years. METHODS: PubMed was searched for papers containing data for novel immunogenic proteins tested on humans till March 2017. Corresponding protein sequences were collected from NCBI and UniProtKB. The set was curated manually for multiple protein fragments, isoforms, and duplicates. RESULTS: The final curated dataset consists of 306 immunogenic proteins tested on humans derived from 47 bacterial microorganisms. Certain proteins have several isoforms. All were considered, and the total protein sequences in the set are 317. The updated set contains 206 new immunogens, compared to the previous VaxiJen bacterial dataset. The average number of immunogens per species is 6.7. The set also contains 12 fusion proteins and 41 peptide fragments and epitopes. The dataset includes the names of bacterial microorganisms, protein names, and protein sequences in FASTA format. CONCLUSION: Currently, the updated VaxiJen bacterial dataset is the best known manually-curated compilation of bacterial immunogens. It is freely available at http://www.ddg-pharmfac.net/vaxi jen/dataset. It can easily be downloaded, searched, and processed. When combined with an appropriate negative dataset, this update could also serve as a training set, allowing enhanced prediction of the potential immunogenicity of unknown protein sequences.

Selection-based design of in silico dengue epitope ensemble vaccines.

Dengue virus affects approximately 130 countries. Twenty-five percentage of infections result in febrile, self-limiting illness; heterotypic infection results in potentially fatal dengue haemorrhagic fever or dengue shock syndrome. Only one vaccine is currently available. Its efficacy is very variable. Thus, to target dengue, we used an innovative immunoinformatics protocol to design a putative epitope ensemble vaccine by selecting an optimal set of highly conserved epitopes with experimentally verified immunogenicity. From 1597 CD4+ and MHC II epitopes, six MHC Class I epitopes (RAVHADMGYW, GPWHLGKLEM, GLYGNGVVTK, NMIIMDEAHF, KTWAYHGSY and WAYHGSYEV) and nine MHC Class II epitopes (LAKAIFKLTYQNKVV, GKIVGLYGNGVVTTS, AAIFMTATPPGSVEA, AAIFMTATPPGTADA, GKTVWFVPSIKAGND, KFWNTTIAVSMANIF, RAIWYMWLGARYLEF, VGTYGLNTFTNMEVQ and WTLMYFHRRDLRLAA) were selected; this candidate vaccine achieved a world population coverage of 92.49%.

In silico prediction of cancer immunogens: current state of the art.

Cancer kills 8 million annually worldwide. Although survival rates in prevalent cancers continue to increase, many cancers have no effective treatment, prompting the search for new and improved protocols. Immunotherapy is a new and exciting addition to the anti-cancer arsenal. The successful and accurate identification of aberrant host proteins acting as antigens for vaccination and immunotherapy is a key aspiration for both experimental and computational research. Here we describe key elements of in silico prediction, including databases of cancer antigens and bleeding-edge methodology for their prediction. We also highlight the role dendritic cell vaccines can play and how they can act as delivery mechanisms for epitope ensemble vaccines. Immunoinformatics can help streamline the discovery and utility of Cancer Immunogens.

In silico design of Mycobacterium tuberculosis epitope ensemble vaccines.

Effective control of Mycobacterium tuberculosis is a global necessity. In 2015, tuberculosis (TB) caused more deaths than HIV. Considering the increasing prevalence of multi-drug resistant forms of M. tuberculosis, the need for effective TB vaccines becomes imperative. Currently, the only licensed TB vaccine is Bacillus Calmette-Guérin (BCG). Yet, BCG has many drawbacks limiting its efficacy and applicability. We applied advanced computational procedures to derive a universal TB vaccine and one targeting East Africa. Our approach selects an optimal set of highly conserved, experimentally validated epitopes, with high projected population coverage (PPC). Through rigorous data analysis, five different potential vaccine combinations were selected each with PPC above 80% for East Africa and above 90% for the World. Two potential vaccines only contained CD8+ epitopes, while the others included both CD4+ and CD8+ epitopes. Our prime vaccine candidate was a putative seven-epitope ensemble comprising: SRGWSLIKSVRLGNA, KPRIITLTMNPALDI, AAHKGLMNIALAISA, FPAGGSTGSL, MLLAVTVSL, QSSFYSDW and KMRCGAPRY, with a 97.4% global PPC and a 92.7% East African PPC.

In silico design of knowledge-based Plasmodium falciparum epitope ensemble vaccines.

Malaria is a global health burden, and a major cause of mortality and morbidity in Africa. Here we designed a putative malaria epitope ensemble vaccine by selecting an optimal set of pathogen epitopes. From the IEDB database, 584 experimentally-verified CD8+ epitopes and 483 experimentally-verified CD4+ epitopes were collected; 89% of which were found in 8 proteins. Using the PVS server, highly conserved epitopes were identified from variability analysis of multiple alignments of Plasmodium falciparum protein sequences. The allele-dependent binding of epitopes was then assessed using IEDB analysis tools, from which the population protection coverage of single and combined epitopes was estimated. Ten conserved epitopes from four well-studied antigens were found to have a coverage of 97.9% of the world population: 7 CD8+ T cell epitopes (LLMDCSGSI, FLIFFDLFLV, LLACAGLAYK, TPYAGEPAPF, LLACAGLAY, SLKKNSRSL, and NEVVVKEEY) and 3 CD4+ T cell epitopes (MRKLAILSVSSFLFV, KSKYKLATSVLAGLL and GLAYKFVVPGAATPYE). The addition of four heteroclitic peptides - single point mutated epitopes - increased HLA binding affinity and raised the predicted world population coverage above 99%.

Designing Epitope Ensemble Vaccines against TB by Selection: Prioritizing Antigens using Predicted Immunogenicity.

Tuberculosis (TB) is a global health burden, and a major cause of mortality and morbidity in West Africa. Here, we select key conserved pathogen epitopes of proven immunogenicity to form a potential TB epitope ensemble vaccine. We compared two vaccine formulations: one comprising class I epitopes from the 13 most prevalent class I epitope-bearing antigens and class II epitopes deriving from the 20 most prevalent class II epitope-bearing antigens and another consisting of epitopes derived solely from 5 antigens identified as the most immunogenic by VaxiJen. In the prevalence analysis, 279 class I and 561 class II epitopes were collected and a subset selected using our published methodology, yielding 32 conserved epitopes. Combining 9 conserved epitopes gave a putative vaccine with predicted population coverage (PPC) over 95%. This consists of ISSGVFLLK, AVAGAAILV, WYYQSGLSI, YQSGLSIVM, MPVGGQSSF, QSSFYSDWY, WDINTPAFEWYYQSGLSIVM, DAPLITNPGGLLEQAVAVEE and NQAVTPAARALPLTSLTSAA. 5 immunogenic antigens VaxiJen-identified yielded 187 epitopes, which we again analyzed using published protocol. This identified 11 conserved epitopes. From this set the highest PPC value (<85%) was obtained by combining: GQQYQAMSAQAAAFH, DDIKATYDKGILTVSVAVSE and AVAGAAILV. We conclude that prioritizing epitope selection using predicted immunogenicity alone is likely to be unduly restrictive and is currently not an optimal or advisable strategy in the design of epitope ensemble vaccines.

Scrutinizing human MHC polymorphism: Supertype analysis using Poisson-Boltzmann electrostatics and clustering.

Peptide-binding MHC proteins are thought the most variable across the human population; the extreme MHC polymorphism observed is functionally important and results from constrained divergent evolution. MHCs have vital functions in immunology and homeostasis: cell surface MHC class I molecules report cell status to CD8+ T cells, NKT cells and NK cells, thus playing key roles in pathogen defence, as well as mediating smell recognition, mate choice, Adverse Drug Reactions, and transplantation rejection. MHC peptide specificity falls into several supertypes exhibiting commonality of binding. It seems likely that other supertypes exist relevant to other functions. Since comprehensive experimental characterization is intractable, structure-based bioinformatics is the only viable solution. We modelled functional MHC proteins by homology and used calculated Poisson-Boltzmann electrostatics projected from the top surface of the MHC as multi-dimensional descriptors, analysing them using state-of-the-art dimensionality reduction techniques and clustering algorithms. We were able to recover the 3 MHC loci as separate clusters and identify clear sub-groups within them, vindicating unequivocally our choice of both data representation and clustering strategy. We expect this approach to make a profound contribution to the study of MHC polymorphism and its functional consequences, and, by extension, other burgeoning structural systems, such as GPCRs.

Protein lipograms.

Linguistic analysis of protein sequences is an underexploited technique. Here, we capitalize on the concept of the lipogram to characterize sequences at the proteome levels. A lipogram is a literary composition which omits one or more letters. A protein lipogram likewise omits one or more types of amino acid. In this article, we establish a usable terminology for the decomposition of a sequence collection in terms of the lipogram. Next, we characterize Uniref50 using a lipogram decomposition. At the global level, protein lipograms exhibit power-law properties. A clear correlation with metabolic cost is seen. Finally, we use the lipogram construction to assign proteomes to the four branches of the tree-of-life: archaea, bacteria, eukaryotes and viruses. We conclude from this pilot study that the lipogram demonstrates considerable potential as an additional tool for sequence analysis and proteome classification.

In Silico Adjuvant Design and Validation.

Adjuvants are substances that boost the protective immune response to vaccine antigens. The majority of known adjuvants have been identified through the use of empirical approaches. Our aim was to identify novel adjuvants with well-defined cellular and molecular mechanisms by combining a knowledge of immunoregulatory mechanisms with an in silico approach. CD4+CD25+FoxP3+ regulatory T cells (Tregs) inhibit the protective immune responses to vaccines by suppressing the activation of antigen presenting cells such as dendritic cells (DCs). In this chapter, we describe the identification and functional validation of small molecule antagonists to CCR4, a chemokine receptor expressed on Tregs. The CCR4 binds the chemokines CCL22 and CCL17 that are produced in large amounts by activated innate cells including DCs. In silico identified small molecule CCR4 antagonists inhibited the migration of Tregs both in vitro and in vivo and when combined with vaccine antigens, significantly enhanced protective immune responses in experimental models.

Towards the knowledge-based design of universal influenza epitope ensemble vaccines.

MOTIVATION: Influenza A viral heterogeneity remains a significant threat due to unpredictable antigenic drift in seasonal influenza and antigenic shifts caused by the emergence of novel subtypes. Annual review of multivalent influenza vaccines targets strains of influenza A and B likely to be predominant in future influenza seasons. This does not induce broad, cross protective immunity against emergent subtypes. Better strategies are needed to prevent future pandemics. Cross-protection can be achieved by activating CD8+ and CD4+ T cells against highly conserved regions of the influenza genome. We combine available experimental data with informatics-based immunological predictions to help design vaccines potentially able to induce cross-protective T-cells against multiple influenza subtypes. RESULTS: To exemplify our approach we designed two epitope ensemble vaccines comprising highly conserved and experimentally verified immunogenic influenza A epitopes as putative non-seasonal influenza vaccines; one specifically targets the US population and the other is a universal vaccine. The USA-specific vaccine comprised 6 CD8+ T cell epitopes (GILGFVFTL, FMYSDFHFI, GMDPRMCSL, SVKEKDMTK, FYIQMCTEL, DTVNRTHQY) and 3 CD4+ epitopes (KGILGFVFTLTVPSE, EYIMKGVYINTALLN, ILGFVFTLTVPSERG). The universal vaccine comprised 8 CD8+ epitopes: (FMYSDFHFI, GILGFVFTL, ILRGSVAHK, FYIQMCTEL, ILKGKFQTA, YYLEKANKI, VSDGGPNLY, YSHGTGTGY) and the same 3 CD4+ epitopes. Our USA-specific vaccine has a population protection coverage (portion of the population potentially responsive to one or more component epitopes of the vaccine, PPC) of over 96 and 95% coverage of observed influenza subtypes. The universal vaccine has a PPC value of over 97 and 88% coverage of observed subtypes. AVAILABILITY AND IMPLEMENTATION: http://imed.med.ucm.es/Tools/episopt.html CONTACT: d.r.flower@aston.ac.uk.

A Cohesive and Integrated Platform for Immunogenicity Prediction.

In silico methods for immunogenicity prediction mine the enormous quantity of data arising from deciphered genomes and proteomes to identify immunogenic proteins. While high and productive immunogenicity is essential for vaccines, therapeutic proteins and monoclonal antibodies should be minimally immunogenic. Here, we present a cohesive platform for immunogenicity and MHC class I and/or II binding affinity prediction. The platform integrates three quasi-independent modular servers: VaxiJen, EpiJen, and EpiTOP. VaxiJen (http://www.ddg-pharmfac.net/vaxijen) predicts immunogenicity of proteins of different origin; EpiJen (http://www.ddg-pharmfac.net/epijen) predicts peptide binding to MHC class I proteins; and EpiTOP (http://www.ddg-pharmfac.net/epitop) predicts peptide binding to MHC class II proteins. The platform is freely accessible and user-friendly. The protocol for immunogenicity prediction is demonstrated by selecting immunogenic proteins from Mycobacterium tuberculosis and predicting how the peptide epitopes within them bind to MHC class I and class II proteins.

Accurate estimation of isoelectric point of protein and peptide based on amino acid sequences.

MOTIVATION: In any macromolecular polyprotic system-for example protein, DNA or RNA-the isoelectric point-commonly referred to as the pI-can be defined as the point of singularity in a titration curve, corresponding to the solution pH value at which the net overall surface charge-and thus the electrophoretic mobility-of the ampholyte sums to zero. Different modern analytical biochemistry and proteomics methods depend on the isoelectric point as a principal feature for protein and peptide characterization. Protein separation by isoelectric point is a critical part of 2-D gel electrophoresis, a key precursor of proteomics, where discrete spots can be digested in-gel, and proteins subsequently identified by analytical mass spectrometry. Peptide fractionation according to their pI is also widely used in current proteomics sample preparation procedures previous to the LC-MS/MS analysis. Therefore accurate theoretical prediction of pI would expedite such analysis. While such pI calculation is widely used, it remains largely untested, motivating our efforts to benchmark pI prediction methods. RESULTS: Using data from the database PIP-DB and one publically available dataset as our reference gold standard, we have undertaken the benchmarking of pI calculation methods. We find that methods vary in their accuracy and are highly sensitive to the choice of basis set. The machine-learning algorithms, especially the SVM-based algorithm, showed a superior performance when studying peptide mixtures. In general, learning-based pI prediction methods (such as Cofactor, SVM and Branca) require a large training dataset and their resulting performance will strongly depend of the quality of that data. In contrast with Iterative methods, machine-learning algorithms have the advantage of being able to add new features to improve the accuracy of prediction. CONTACT: yperez@ebi.ac.uk AVAILABILITY AND IMPLEMENTATION: The software and data are freely available at https://github.com/ypriverol/pIRSupplementary information: Supplementary data are available at Bioinformatics online.