StarGazer: A Hybrid Intelligence Platform for Drug Target Prioritization and Digital Drug Repositioning Using Streamlit.

Target prioritization is essential for drug discovery and repositioning. Applying computational methods to analyze and process multi-omics data to find new drug targets is a practical approach for achieving this. Despite an increasing number of methods for generating datasets such as genomics, phenomics, and proteomics, attempts to integrate and mine such datasets remain limited in scope. Developing hybrid intelligence solutions that combine human intelligence in the scientific domain and disease biology with the ability to mine multiple databases simultaneously may help augment drug target discovery and identify novel drug-indication associations. We believe that integrating different data sources using a singular numerical scoring system in a hybrid intelligent framework could help to bridge these different omics layers and facilitate rapid drug target prioritization for studies in drug discovery, development or repositioning. Herein, we describe our prototype of the StarGazer pipeline which combines multi-source, multi-omics data with a novel target prioritization scoring system in an interactive Python-based Streamlit dashboard. StarGazer displays target prioritization scores for genes associated with 1844 phenotypic traits, and is available via https://github.com/AstraZeneca/StarGazer.

Knowledge graph-based recommendation framework identifies drivers of resistance in EGFR mutant non-small cell lung cancer.

Resistance to EGFR inhibitors (EGFRi) presents a major obstacle in treating non-small cell lung cancer (NSCLC). One of the most exciting new ways to find potential resistance markers involves running functional genetic screens, such as CRISPR, followed by manual triage of significantly enriched genes. This triage process to identify 'high value' hits resulting from the CRISPR screen involves manual curation that requires specialized knowledge and can take even experts several months to comprehensively complete. To find key drivers of resistance faster we build a recommendation system on top of a heterogeneous biomedical knowledge graph integrating pre-clinical, clinical, and literature evidence. The recommender system ranks genes based on trade-offs between diverse types of evidence linking them to potential mechanisms of EGFRi resistance. This unbiased approach identifies 57 resistance markers from >3,000 genes, reducing hit identification time from months to minutes. In addition to reproducing known resistance markers, our method identifies previously unexplored resistance mechanisms that we prospectively validate.

An open approach to systematically prioritize causal variants and genes at all published human GWAS trait-associated loci.

Genome-wide association studies (GWASs) have identified many variants associated with complex traits, but identifying the causal gene(s) is a major challenge. In the present study, we present an open resource that provides systematic fine mapping and gene prioritization across 133,441 published human GWAS loci. We integrate genetics (GWAS Catalog and UK Biobank) with transcriptomic, proteomic and epigenomic data, including systematic disease-disease and disease-molecular trait colocalization results across 92 cell types and tissues. We identify 729 loci fine mapped to a single-coding causal variant and colocalized with a single gene. We trained a machine-learning model using the fine-mapped genetics and functional genomics data and 445 gold-standard curated GWAS loci to distinguish causal genes from neighboring genes, outperforming a naive distance-based model. Our prioritized genes were enriched for known approved drug targets (odds ratio = 8.1, 95% confidence interval = 5.7, 11.5). These results are publicly available through a web portal ( http://genetics.opentargets.org ), enabling users to easily prioritize genes at disease-associated loci and assess their potential as drug targets.

Open Targets Platform: supporting systematic drug-target identification and prioritisation.

The Open Targets Platform (https://www.targetvalidation.org/) provides users with a queryable knowledgebase and user interface to aid systematic target identification and prioritisation for drug discovery based upon underlying evidence. It is publicly available and the underlying code is open source. Since our last update two years ago, we have had 10 releases to maintain and continuously improve evidence for target-disease relationships from 20 different data sources. In addition, we have integrated new evidence from key datasets, including prioritised targets identified from genome-wide CRISPR knockout screens in 300 cancer models (Project Score), and GWAS/UK BioBank statistical genetic analysis evidence from the Open Targets Genetics Portal. We have evolved our evidence scoring framework to improve target identification. To aid the prioritisation of targets and inform on the potential impact of modulating a given target, we have added evaluation of post-marketing adverse drug reactions and new curated information on target tractability and safety. We have also developed the user interface and backend technologies to improve performance and usability. In this article, we describe the latest enhancements to the Platform, to address the fundamental challenge that developing effective and safe drugs is difficult and expensive.

Open Targets Genetics: systematic identification of trait-associated genes using large-scale genetics and functional genomics.

Open Targets Genetics (https://genetics.opentargets.org) is an open-access integrative resource that aggregates human GWAS and functional genomics data including gene expression, protein abundance, chromatin interaction and conformation data from a wide range of cell types and tissues to make robust connections between GWAS-associated loci, variants and likely causal genes. This enables systematic identification and prioritisation of likely causal variants and genes across all published trait-associated loci. In this paper, we describe the public resources we aggregate, the technology and analyses we use, and the functionality that the portal offers. Open Targets Genetics can be searched by variant, gene or study/phenotype. It offers tools that enable users to prioritise causal variants and genes at disease-associated loci and access systematic cross-disease and disease-molecular trait colocalization analysis across 92 cell types and tissues including the eQTL Catalogue. Data visualizations such as Manhattan-like plots, regional plots, credible sets overlap between studies and PheWAS plots enable users to explore GWAS signals in depth. The integrated data is made available through the web portal, for bulk download and via a GraphQL API, and the software is open source. Applications of this integrated data include identification of novel targets for drug discovery and drug repurposing.

Open Targets Platform: new developments and updates two years on.

The Open Targets Platform integrates evidence from genetics, genomics, transcriptomics, drugs, animal models and scientific literature to score and rank target-disease associations for drug target identification. The associations are displayed in an intuitive user interface (https://www.targetvalidation.org), and are available through a REST-API (https://api.opentargets.io/v3/platform/docs/swagger-ui) and a bulk download (https://www.targetvalidation.org/downloads/data). In addition to target-disease associations, we also aggregate and display data at the target and disease levels to aid target prioritisation. Since our first publication two years ago, we have made eight releases, added new data sources for target-disease associations, started including causal genetic variants from non genome-wide targeted arrays, added new target and disease annotations, launched new visualisations and improved existing ones and released a new web tool for batch search of up to 200 targets. We have a new URL for the Open Targets Platform REST-API, new REST endpoints and also removed the need for authorisation for API fair use. Here, we present the latest developments of the Open Targets Platform, expanding the evidence and target-disease associations with new and improved data sources, refining data quality, enhancing website usability, and increasing our user base with our training workshops, user support, social media and bioinformatics forum engagement.

Open Targets: a platform for therapeutic target identification and validation.

We have designed and developed a data integration and visualization platform that provides evidence about the association of known and potential drug targets with diseases. The platform is designed to support identification and prioritization of biological targets for follow-up. Each drug target is linked to a disease using integrated genome-wide data from a broad range of data sources. The platform provides either a target-centric workflow to identify diseases that may be associated with a specific target, or a disease-centric workflow to identify targets that may be associated with a specific disease. Users can easily transition between these target- and disease-centric workflows. The Open Targets Validation Platform is accessible at https://www.targetvalidation.org.

Non-invasive mapping of the gastrointestinal microbiota identifies children with inflammatory bowel disease.

BACKGROUND: Pediatric inflammatory bowel disease (IBD) is challenging to diagnose because of the non-specificity of symptoms; an unequivocal diagnosis can only be made using colonoscopy, which clinicians are reluctant to recommend for children. Diagnosis of pediatric IBD is therefore frequently delayed, leading to inappropriate treatment plans and poor outcomes. We investigated the use of 16S rRNA sequencing of fecal samples and new analytical methods to assess differences in the microbiota of children with IBD and other gastrointestinal disorders. METHODOLOGY/PRINCIPAL FINDINGS: We applied synthetic learning in microbial ecology (SLiME) analysis to 16S sequencing data obtained from i) published surveys of microbiota diversity in IBD and ii) fecal samples from 91 children and young adults who were treated in the gastroenterology program of Children's Hospital (Boston, USA). The developed method accurately distinguished control samples from those of patients with IBD; the area under the receiver-operating-characteristic curve (AUC) value was 0.83 (corresponding to 80.3% sensitivity and 69.7% specificity at a set threshold). The accuracy was maintained among data sets collected by different sampling and sequencing methods. The method identified taxa associated with disease states and distinguished patients with Crohn's disease from those with ulcerative colitis with reasonable accuracy. The findings were validated using samples from an additional group of 68 patients; the validation test identified patients with IBD with an AUC value of 0.84 (e.g. 92% sensitivity, 58.5% specificity). CONCLUSIONS/SIGNIFICANCE: Microbiome-based diagnostics can distinguish pediatric patients with IBD from patients with similar symptoms. Although this test can not replace endoscopy and histological examination as diagnostic tools, classification based on microbial diversity is an effective complementary technique for IBD detection in pediatric patients.

Characterisation of the Trichinella spiralis deubiquitinating enzyme, TsUCH37, an evolutionarily conserved proteasome interaction partner.

BACKGROUND: Trichinella spiralis is a zoonotic parasitic nematode that causes trichinellosis, a disease that has been identified on all continents except Antarctica. During chronic infection, T. spiralis larvae infect skeletal myofibres, severely disrupting their differentiation state. METHODOLOGY AND RESULTS: An activity-based probe, HA-Ub-VME, was used to identify deubiquitinating enzyme (DUB) activity in lysate of T. spiralis L1 larvae. Results were analysed by immuno-blot and immuno-precipitation, identifying a number of potential DUBs. Immuno-precipitated proteins were subjected to LC/MS/MS, yielding peptides with sequence homology to 5 conserved human DUBs: UCH-L5, UCH-L3, HAUSP, OTU 6B and Ataxin-3. The predicted gene encoding the putative UCH-L5 homologue, TsUCH37, was cloned and recombinant protein was expressed and purified. The deubiquitinating activity of this enzyme was verified by Ub-AMC assay. Co-precipitation of recombinant TsUCH37 showed that the protein associates with putative T. spiralis proteasome components, including the yeast Rpn13 homologue ADRM1. In addition, the UCH inhibitor LDN-57444 exhibited specific inhibition of recombinant TsUCH37 and reduced the viability of cultured L1 larvae. CONCLUSIONS: This study reports the identification of the first T. spiralis DUB, a cysteine protease that is putatively orthologous to the human protein, hUCH-L5. Results suggest that the interaction of this protein with the proteasome has been conserved throughout evolution. We show potential for the use of inhibitor compounds to elucidate the role of UCH enzymes in T. spiralis infection and their investigation as therapeutic targets for trichinellosis.

The tetraspanin CD82 is specifically recruited to fungal and bacterial phagosomes prior to acidification.

CD82 is a member of the tetraspanin superfamily, whose physiological role is best described in the context of cancer metastasis. However, CD82 also associates with components of the class II major histocompatibility complex (MHC) antigen presentation pathway, including class II MHC molecules and the peptide-loading machinery, as well as CD63, another tetraspanin, suggesting a role for CD82 in antigen presentation. Here, we observe the dynamic rearrangement of CD82 after pathogen uptake by imaging CD82-mRFP1 expressed in primary living dendritic cells. CD82 showed rapid and specific recruitment to Cryptococcus neoformans-containing phagosomes compared to polystyrene-containing phagosomes, similar to CD63. CD82 was also actively recruited to phagosomes containing other pathogenic fungi, including Candida albicans and Aspergillus fumigatus. Recruitment of CD82 to fungal phagosomes occurred independently of Toll-like receptor (TLR) signaling. Recruitment was not limited to fungi, as bacterial organisms, including Escherichia coli and Staphylococcus aureus, also induced CD82 recruitment to the phagosome. CD82 intersected the endocytic pathway used by lipopolysaccharide (LPS), implicating CD82 in trafficking of small, pathogen-associated molecules. Despite its partial overlap with lysosomal compartments, CD82 recruitment to C. neoformans-containing phagosomes occurred independently of phagosome acidification. Kinetic analysis of fluorescence imaging revealed that CD82 and class II MHC simultaneously appear in the phagosome, indicating that the two proteins may be associated. Together, these data show that the CD82 tetraspanin is specifically recruited to pathogen-containing phagosomes prior to fusion with lysosomes.

Screening individual hybridomas by microengraving to discover monoclonal antibodies.

The demand for monoclonal antibodies (mAbs) in biomedical research is significant, but the current methodologies used to discover them are both lengthy and costly. Consequently, the diversity of antibodies available for any particular antigen remains limited. Microengraving is a soft lithographic technique that provides a rapid and efficient alternative for discovering new mAbs. This protocol describes how to use microengraving to screen mouse hybridomas to establish new cell lines producing unique mAbs. Single cells from a polyclonal population are isolated into an array of microscale wells (approximately 10(5) cells per screen). The array is then used to print a protein microarray, where each element contains the antibodies captured from individual wells. The antibodies on the microarray are screened with antigens of interest, and mapped to the corresponding cells, which are then recovered from their microwells by micromanipulation. Screening and retrieval require approximately 1-3 d (9-12 d including the steps for preparing arrays of microwells).

Optimization of the surfaces used to capture antibodies from single hybridomas reduces the time required for microengraving.

The most common method for the generation of monoclonal antibodies involves the identification and isolation of hybridomas from polyclonal populations. The discovery of new antibodies for biochemical and immunohistochemical assays in a rapid and efficient manner, however, remains a challenge. Here, a series of experiments are described that realize significant improvements to an approach for screening large numbers of single cells to identify antigen-specific monoclonal antibodies in a high-throughput manner (10(5)-10(6) cells in less than 12 h). The soft lithographic process called microengraving yields microarrays of monoclonal antibodies that can be correlated to individual hybridomas; the cells can then be retrieved and expanded to establish new cell lines. The factors examined here included the glass slide used for the microarray, the buffer used to deposit capture antibodies onto the glass, the type of polyclonal antibodies used to capture the secreted antibodies, and the time required for microengraving. Compared to earlier reports of this method, these studies resulted in increased signal-to-noise ratios for individual elements in the microarrays produced, and a considerable decrease in the time required to produce one microarray from a set of cells (from 2-4 h to 3-10 min). These technical advances will improve the throughput and reduce the costs for this alternative to traditional screening by limiting serial dilution.

Profiling antibody responses by multiparametric analysis of primary B cells.

Determining the efficacy of a vaccine generally relies on measuring neutralizing antibodies in sera. This measure cannot elucidate the mechanisms responsible for the development of immunological memory at the cellular level, however. Quantitative profiles that detail the cellular origin, extent, and diversity of the humoral (antibody-based) immune response would improve both the assessment and development of vaccines. Here, we describe a novel approach to collect multiparametric datasets that describe the specificity, isotype, and apparent affinity of the antibodies secreted from large numbers of individual primary B cells (approximately 10(3)-10(4)). The antibody/antigen binding curves obtained by this approach can be used to classify closely related populations of cells using algorithms for data clustering, and the relationships among populations can be visualized graphically using affinity heatmaps. The technique described was used to evaluate the diversity of antigen-specific antibody-secreting cells generated during an in vivo humoral response to a series of immunizations designed to mimic a multipart vaccination. Profiles correlating primary antibody-producing cells with the molecular characteristics of their secreted antibodies should facilitate both the evaluation of candidate vaccines and, broadly, studies on the repertoires of antibodies generated in response to infectious or autoimmune diseases.