Select Better Antibodies Using NGS Data

Simply select better antibodies with the IGX Platform 1 Simply select better antibodies with the IGX Platform WHITE PAPER Simply select better antibodies with the IGX Platform INTRODUCTION 1 Discovery Workflow 2 Challenges 4 THE IGX PLATFORM 6 Modular 7 Core Apps 10 Antibody Discovery Module 12 Antibody Diversity 13 Structural Modeling 14 Tracking Enrichment 16 NGS APPLICATION USE CASES 17 Hit Selection 18 Hit Expansion 20 Enrichment Analysis 22 MANAGE 24 API 25 Integrate 26 Find 29 Collect 29 Share 30 Secure 31 Technical Assistance 32 SUMMARY 33 REFERENCES 34 TABLE OF CONTENTS Simply select better antibodies with the IGX Platform Simply select better antibodies with the IGX Platform 1 INTRODUCTION The field of antibody therapeutics is experiencing explosive growth. Antibody development has proven to yield effective drug molecules and the success stories are coming in: from over 500 clinical studies, the success rate between phase I and approval lies between 17 and 25% [1], which is double the average general approval rate (9,6% between 2006-2015 [2]). A variety of different antibody discovery platforms are in use today. They rely on the generation of thousands to millions of distinct candidates from which a small selection with the best therapeutic characteristics must be picked. Although considerable improvements have been made to the throughput of biochemical and antibody binding assays, the current rate of discovery is still limited by the in vitro screening phase of the antibody discovery workflow [3]. Moreover, as the path to clinical approval of leads requires a significant investment; there is a need for dropping unfit candidates during early discovery to lower investment cost and maximize the chances of bringing a successful therapeutic to the market. Lastly, obtaining a high number of binders against a target is crucial to broadly protect the IP of the lead therapeutic, and secure a return on investment. Next-Generation Sequencing (NGS) data captured throughout the antibody discovery process has shown to mitigate the aforementioned limitations and maximize the chances to bring viable therapeutics to the clinic and is therefore playing an increasingly important role in all discovery workflows. By tightly integrating sequences with lab measurements, antibody candidates can be analyzed in greater numbers and detail than ever before. However, the analysis of high-throughput immune repertoire sequencing (Rep-Seq) data can be challenging, as it requires expertise in different fields, such as immunology and bioinformatics, and a powerful IT infrastructure. Together with leading experts in the fields of antibody discovery and vaccine development, ENPICOM has developed the Antibody Discovery Module – a suite of IGX Platform Apps specifically tailored to address the unique data analysis challenges faced in the field of antibody discovery. With the Antibody Discovery Module, the IGX Platform empowers life scientists to independently perform complex analyses, identify and discard irrelevant candidates in early discovery, and select the best antibody candidates. Simply select better antibodies with the IGX Platform 2 DISCOVERY WORKFLOW Table 1. Rep-Seq application in antibody discovery Discovery workflow Starting material Sequencing technologies In-vitro screening throughput Liability occurence Main NGS application Hybridoma Immunized animal Sanger Bulk NGS Hundreds Low Hit expansion Single B Cell Immunized/healthy repertoire Single-cell NGS Bulk NGS Thousands Medium Hit selection Hit expansion Display technology Immunized/healthy repertoire, synthetic library Sanger Bulk NGS Millions Medium/Low Library QC Hit selection Hit expansion LEVERAGING REP-SEQ NGS technologies provide an unparalleled level of sequencing depth that, for the first time, allows for the identification of millions of receptor sequences in a single experiment. The increased depth significantly boosts the available candidate space from which novel antibody sequences can be recovered. Each discovery platform leverages highthroughput sequencing data differently to improve and accelerate the discovery process (Table 1). Simply select better antibodies with the IGX Platform 3 Hybridoma The production of immortalized B cell lines (hybridomas) is labor-intensive and typically only hundreds of antibodies can be tested for binding, leaving a large part of the candidate space unexplored [4,5]. NGS is increasingly applied to expand the candidate output (socalled ‘hit expansion’). Millions of receptor sequences can be identified through highthroughput sequencing and clustered together with a small set of validated binders (typically Sanger sequences). The enlarged and diversified candidate scope increases the chance of successfully identifying a suitable candidate and can be used to discover candidates with more desirable features (e.g., increased affinity or stability, fewer sequence liabilities). Single B cell The identification of paired heavy and light chains in single B cell workflows is frequently done with plate-based NGS or high-throughput single-cell sequencing technologies, and typically yields hundreds to thousands of distinct antibody variants [6]. In contrast to display and hybridoma workflows, the information associated with the sequences is usually limited to antigen-specificity. Since phenotypical information is minimal, antibody selection is based on clone abundance, clonal overlap, and sequence properties from single-cell NGS data. The recent development of high-throughput B cell screening technologies enables more extensive phenotypical characterization, but sequence information can still be crucial for candidate selection. Moreover, bulk NGS data can provide a much larger pool of unpaired sequences, which is valuable for hit expansion and to increase the accuracy of clone abundance estimations. Display technology The display of antibody fragments on phages, cells and other carriers, and subsequent selection through panning rounds, enables the screening of millions of distinct receptor variants in a single experiment [7]. Display workflows can make use of NGS data for quality control of the library (e.g., to ensure library diversity and avoid biases) and to analyze clone enrichment profiles throughout panning rounds. Promising candidates can be identified by studying their sequence enrichment across panning rounds, rather than relying solely on the abundance of candidates in the output of the final panning round. Furthermore, by including different panning conditions and a variety of antigens (e.g., targets, counter-targets, and irrelevant antigens), a robust candidate pool with diverse binding profiles can be obtained. Simply select better antibodies with the IGX Platform 4 Antibody discovery workflows typically yield data output of different types, including receptor repertoire sequences and corresponding output from assays and other wet-lab assessments. To accurately characterize antibody candidates, tight integration of sequencing and experimental data is key. The integrated analysis requires expertise in different fields, including immunology and bioinformatics, to tackle the main challenges: Accurate clone annotation Antibodies are the product of gene rearrangements and contain non-templated sequence elements. Consequently, there is no reference database with full receptor sequences, but only for the gene segments that are rearranged (e.g., the V, D, and J segments). Affinity maturation increases the diversity from germline even further. This brings challenges to the detection and correction of PCR and sequencing errors in Rep-Seq data and requires specific techniques to process and annotate such data. Error correction NGS technologies allow for an unparalleled exploration of immune repertoires. The sequencing depth and throughput of NGS comes with significant levels of sequencing and PCR errors that need to be accounted for in the analysis. Unique Molecular Identifiers (UMI) allow to correct for this by linking reads to original molecules and remove errors and amplification biases. Error correction using UMIs requires sophisticated algorithms and quality control metrics to optimize the library preparation protocol. Pair chain sequencing processing Most B and T cell receptors are formed by two chains and their binding properties can thus not be fully defined by single chain sequencing alone. Novel library preparation technologies have allowed to keep a link between these by barcoding the receptor mRNA or by indexing sorted cells. During the annotation of sequencing data these cell identifiers need to be tracked to keep the link between two chains of a receptor. To maximize insights, this chain linkage needs to be maintained throughout every downstream analysis. Data integration Integrating data from different sequencing technologies requires a flexible annotation and analysis setup. Integrating clinical and/ or experimental metadata with sequencing information often becomes complicated without the right analysis tool. Moreover, the disconnect between systems like ELN/LIMS adds to the challenge. Having a central source for accurate data helps ensure consistency, increase efficiency, and reduce mistakes. CHALLENGES IN DATA ANALYTICS FOR ANTIBODY DISCOVERY Simply select better antibodies with the IGX Platform 5 Large-scale repertoire interpretation Two common types of analysis in antibody discovery workflows are: (1) the clustering of sequences into groups that (presumably) share properties such as binding specificity [8], and (2) phylogenetic analysis (i.e., ‘lineage tracing’). Given the high volume of NGS data, the computational requirements for these analyses can quickly become a bottleneck and requires technical expertise to run efficiently. Analysis of multi-layered datasets Cluster visualizations and phylogenetic trees of NGS repertoires are often static and unwieldy, making them complicated to navigate. Antibody discovery campaigns produce very diverse sequence-associated metadata and assay data that can be crucial to select the right candidates. Dynamic and interactive visualizations make it easy to overlay and visualize many layers of data at once and simplify the identification of the right candidates. Complex panning designs Antibody discovery campaigns using display technologies can generate numerous datasets coming from various panning rounds and conditions. These datasets are used to identify clone enrichment and specificity against different antigens (e.g., target, countertargets, and irrelevant antigens). Complex analysis workflows – with operations such as intersections, unions, or subtractions – that match the experimental design are needed to fully take advantage of the powerful combination of NGS and antibody panning. De-risking antibody development Sequences often contain unwanted motifs that, when exposed at the surface of the molecule, impact the (clinical) developability of an antibody. To accurately identify these sequence liabilities, the linear sequence is insufficient, as the 3D structure of the antibody will determine the motif impact. While requiring significant computational resources, de novo structural modeling provides detailed insight into the surface exposure of unwanted motifs, leading to highly reliable liability predictions. Simply select better antibodies with the IGX Platform 6 TO SELECT THE BEST CANDIDATES The IGX Platform is an end-to-end cloud solution that removes the complexity of managing and analyzing Rep-Seq data. It supports B and T cells, Sanger and NGS data, paired and single-chain receptor sequences, and has a flexible setup to integrate experimental (assay) results, clinical variables, and in silico predicted properties. By integrating all information in a single environment, it enables scientists to analyze their Rep-Seq data from all angles. This makes it a powerful solution for organizations developing antibodies, vaccines, and other immune-modulating drugs, and for researchers studying the immune system and its related diseases. Securely operating in the cloud, specialized Apps provide intuitive access to extensive analysis and visualization options. The IGX Platform gives every scientist the independence to analyze the adaptive immune repertoire, regardless of their bioinformatics skills. It takes care of data integration and harmonization, so the focus can be on the science, not the coding. The IGX Platform ensures researchers will not miss an insight and select the best candidates for further development. THE IGX PLATFORM Simply select better antibodies with the IGX Platform 7 AND SCALABLE ARCHITECTURE The IGX Platform offers a suite of highly specialized Apps (Figure 1). Users can pick and choose Apps and modules that are most suited to their specific workflow. This provides an easy and cost-effective way for each customer to access the features and analysis methods tailored to their needs. Our Apps are developed and improved in close collaboration with industry and academia, to make sure they provide solutions for a broad variety of workflows that fit the researchers’ needs. Each App offers highly specific analyses and visualizations options that can accomplish a variety of tasks: e.g., performing quality control of NGS data, tracking enrichment, analyzing gene usage, computing phylogenetic trees, or predicting liabilities for further development. The IGX Platform automatically scales its resources, which ensures maximal responsiveness and performance. Resources can be scaled up (more computational power per node) and scaled out (larger number of computational nodes) independently, as required. Apps can also be scaled as needed, independently of other Apps and the platform itself, to ensure top-notch performance when analyzing large datasets. MODULAR Simply select better antibodies with the IGX Platform 8 Figure 1. Overview of Apps and features of the IGX Platform. Simply select better antibodies with the IGX Platform 9 Our Apps are developed in close collaboration with industry and academic partners, to make sure they provide solutions for a broad variety of workflows. Simply select better antibodies with the IGX Platform 10 CORE APPS TO COMPOSE YOUR PERFECT WORKFLOW Enables quality control of FASTQ files and profiled samples. Different interactive visualizations are available, depending on the data type. Read- and sample-level quality statistics are shown in interactive visualizations, allowing users to easily identify batch effects and perform thorough quality inspection on samples. FEATURES ✓ FASTQ QC: read length distribution, per base/sample quality ✓ Read fate of profiled sequences Offers a scalable and intuitive way to go from raw sequencing data to annotated clones. It annotates B and T cell receptors from multiple species sequenced with any technology. Sequence templates can be configured to process any amplicon structure, while quality control (QC) templates allow extensive finetuning of the criteria for accepting clones. It supports plate- and microfluidics-based protocols (including 10x Genomics), linking paired chains using barcodes and UMI error correction to ensure data of the highest quality. FEATURES ✓ Full annotation of immunoglobulin regions ✓ Sequencing technology agnostic (FASTQ, FASTA, and AB1) ✓ Flexible templates for sequence structure and QC ✓ Streamlined batch sample processing ✓ Comprehensive UMI support ✓ Paired-chain processing through barcodes and linkers The IGX Platform offers set of core Apps that form the foundation of the platform and provide a solid basis for all end-to-end analysis workflows. Generates intuitive visualizations to provide insight into the dynamics of the antibody response. It includes several analyses with interactive plots: V/J Gene Usage, Diversity Measures, and CDR3 Length Distribution. Tight metadata integration enables easy data stratification, while toggles, sliders, and tooltips provide intuitive control over plots. FEATURES ✓ V/J Gene Usage, Diversity Measures & CDR3 Length Distribution plots ✓ Intuitive plotting controls ✓ Stratification using metadata Simply select better antibodies with the IGX Platform 11 Figure 2. Example Single Cell NGS protocol on the IGX-Profile Sequence Template Builder. Simply select better antibodies with the IGX Platform 12 TO SIMPLY SELECT BETTER ANTIBODIES The Antibody Discovery Module is a set of specialized Apps specifically engineered to identify better and more diverse antibody candidates faster. Fueled by the powerful data management functionalities of the IGX Platform, our Antibody Discovery Module helps you expand and diversify your candidate pool, whether you are using NGS, Sanger, single B cells, or a combination of these technologies. Efficiently integrate sequence and experimental assay data or in-silico predictions into your analysis workflows and perform cluster, phylogenetic and display enrichment analyses in a secure and scalable environment. To ensure the best antibody candidates are selected for follow-up studies the Antibody Discovery Module also enables high-throughput structural modeling of antibodies to accurately identify and annotate exposed liabilities. This makes it a powerful solution to streamline any antibody discovery workflow and easily identify the best antibody candidates for further development. ANTIBODY DISCOVERY MODULE Simply select better antibodies with the IGX Platform 13 Antibody discovery campaigns are getting more complex and often include datasets from different origins and sequencing technologies. These datasets generally consist of sequences of antigen-specific antibodies (e.g., Sanger or single-cell sequencing) and Rep-Seq of unknown specificity (e.g., unsorted NGS). Finding antigen-specific antibodies with the best characteristics across different datasets requires two steps. First, sequences of different datasets need to be grouped together based on a feature or sequence similarity of a region of interest. Second, groups of interested needs to be explored further to find the best candidates for further development. These steps are commonly followed in workflows like hit selection and expansion. The IGX Platform provides two complementary Apps to aid in such analyses. Identifies sequence clusters based on userdefined parameters such as heavy and/or light sequence similarity percentage, gene usage, and CDR3 length. It builds clusters based on the similarity of any region of the antibody sequence, as specified by the user (e.g., CDR3, CDR1+2+3, or full VDJ). IGX-Cluster can process together sequences originating from any data sources supported by the platform. Clustering across samples from different sequencing technologies simplifies the comparison of antibody candidates from different experiments. FEATURES ✓ Scalable computations ✓ Multi-sample support ✓ Flexible multiparametric cluster definitions Is a two-stage rocket. First, clusters generated by IGX-Cluster are prioritized based on properties and content (e.g., sample origin, SHM levels, or absence of specific liabilities) and phylogenetic trees are computed for clusters of interest. Second, powerful interactive tree visualizations are generated which link visual variables such as color, shape, and size to available metadata. Lineages are easily visualized, and interesting metadata, such as affinity measurements, can be overlaid onto the phylogenetic tree to facilitate candidate selection. FEATURES ✓ Comprehensive cluster overview ✓ Interactive phylogenetic trees ✓ Metadata overlays ✓ Multiple sequence alignment of selected clones DIVE INTO THE REPERTOIRE ANTIBODY DIVERSITY Simply select better antibodies with the IGX Platform 14 Therapeutic antibodies need to be expressed in large volumes to be clinically viable. Antibody engineering can improve expression levels although this is an expensive and timeconsuming process. To avoid unnecessary engineering of antibody candidates, liabilities that reduce expression levels can be predicted from the sequence. The identification of sequence liabilities can help to only select and express candidates that require less modifications to improve their expression yields. 3D-model based liability predictions are far superior to motif searches because they account for motifs exposure in the antibody surface. The thoroughly validated SAbPred algorithm generates 3D-model-based liability prediction and developability assessment of antibody candidates. IGX-Annotate enables high-throughput execution of the SAbPred algorithms and integration of the results in the discovery process. DE-RISK YOUR ANTIBODY SELECTION STRUCTURAL MODELING Is a powerful App that enables high-throughput structural modeling of antibodies to accurately identify and annotate exposed liabilities. In addition, it provides an assessment of the developability of antibody candidates by comparing the developability metrics of identified candidates to those of all clinically approved antibodies on the market. Exposed liabilities and developability metrics are scored using a customizable penalty system, and all information is directly linked to the antibody candidates using tags. These tags can be overlaid onto phylogenetic trees, enrichment plots or used for prioritization in the Basket environment, making it easy to select those ‘needle-in-a-haystack’ candidates. FEATURES ✓ High-throughput structural modeling ✓ Accurate prediction of exposed liabilities ✓ Developability assessments (based on clinically approved monoclonal antibodies) Simply select better antibodies with the IGX Platform 15 Figure 3. IGX-Annotate liability score configuration. Simply select better antibodies with the IGX Platform 16 Provides a fast, flexible way to discover a diverse set of highly enriched candidates from display screening. By allowing users to freely join, intersect and subtract datasets, its visual workflow builder supports the analysis of any panning campaign design, no matter the complexity. For example, antibodies binding irrelevant antigens in control panning conditions can be easily subtracted to increase candidate selection precision. Results can be explored through interactive tables and graphs, where any metadata or assay data can be integrated and combined with powerful filtering and sorting options to select the most promising antibody candidates. User-defined representatives of the groups of interest can be easily extracted for further study. FEATURES ✓ Flexible analysis operations (join, intersect, and subtract datasets) ✓ Interactive enrichment visualizations ✓ Quick extraction of user-defined enriched group representatives ✓ Metadata overlays IDENTIFY ENRICHED CANDIDATES Display technologies screen antibody libraries against one or more target antigens during several panning rounds. Antibody candidates are selected after these panning rounds to find strong binders to the targets of interest. However, this process is not failproof: parameters like initial clonal frequency in the antibody library and screening conditions can create biases and prevent the identification of the best candidates. Moreover, panning rounds against counter, or irrelevant targets can be extremely useful to discard un-specific or “sticky” antibodies, but the data analysis requires advanced methods. IGX-Track has been built from the ground up to support the data analysis of complex enrichment strategies, offers easy visualization of any metadata, and integrates liability predictions directly into the App. TRACKING ENRICHMENT Simply select better antibodies with the IGX Platform 17 NGS APPLICATION USE CASES Figure 4. IGX-Track results can be viewed in tables and graphs, filtered and sorted based on enrichment between panning rounds and any other available metadata. Simply select better antibodies with the IGX Platform 18 The repertoire of antibodies generated against an antigen is highly diverse, especially after affinity maturation. Low-throughput screening methods typically allow for an in-depth characterization of only a small portion of the available candidate pool. High-throughput screening methods capture a larger portion, but with less granularity and sensitivity. Single B cell and display technology discovery methods utilize FACS and panning rounds, respectively, to select antibodies. These methods rely on candidates binding a target antigen and thereby give some indication of the potential relevance of the antibody. In immunization-based discovery strategies, repertoire characteristics that can inform on functional properties of the antibodies, like clonal expansion, are indicative of antigen affinity [9]. In both cases, high-throughput sequencing allows to identify many antibody candidates at the same time and gather important characteristics that enable the selection of a diverse set of candidates with high potential. The IGX Platform enables you to select the most promising antibody candidates from the NGS repertoire data by: ✓ Processing UMIs during clone annotation to correct sequence errors and assess the clonal abundance with higher accuracy ✓ Clustering antibody sequences based on VDJ or CDR3 similarities from different samples regardless of the sequencing technology (e.g., panning rounds output together or single-cell sequencing data with bulk repertoire) ✓ Using high-throughput structural modeling to confidently predict sequence liabilities and assess the developability of candidates ✓ Exploring antibody clusters of interest by their size (indicative of clonal expansion), presence of confirmed binders, or any other experimental information (Figure 5) ✓ Extracting representative antibodies of clusters of interest for expression and characterization. NGS data provides orders of magnitude more data compared to low-throughput screen methods, vastly increasing the candidate pool. When combined with a tailored data analysis solution, it allows for precision and sensitivity when selecting antibodies to express and analyze in the lab. HIT SELECTION Simply select better antibodies with the IGX Platform 19 Figure 5. The Cluster Overview in IGX-Branch enables users to select the most promising antibody clusters, by making use of the flexible metadata filters and overlays. Simply select better antibodies with the IGX Platform 20 Traditionally, hybridoma discovery workflows rely on the expression of a limited number of antibody candidates for testing with laboratory assays. This approach provides a small selection of well-characterized antibodies but unavoidably misses many antibodies with potentially better properties. Antibodies not captured in the hybridomas with similar sequences to known binders might have less developmental liabilities or higher somatic hypermutation (SHM) rates, which has been reported to correlate with stronger binding and better properties [10]. Integrating these well-characterized antibody sequences into an analysis with larger NGS enables researchers to expand and diversify the candidate pool and find better antibodies that were missed upon selection. Hit expansion with NGS immune repertoire samples can be easily performed in the IGX Platform by: ✓ Profiling any NGS or paired chain data (using barcodes, linkers or clone identifiers) and annotating the affinity maturation (SHM) in all antibody regions ✓ Linking the assay data (metadata) of the characterized antibodies with their sequences with a simple Excel table ✓ Clustering bulk NGS or single-cell data with Sanger sequences of characterized antibodies based on sequence or paratope homology and other immunologically relevant features ✓ Selecting the clusters where clusters where known binders are present and prioritizing them with the metadata linked to the characterized sequences ✓ Using high-throughput structural modeling to predict exposed sequence liabilities and assess the developability of candidates with confidence and ease ✓ Exploring the lineages within the cluster of interest to identify the antibody candidates with desired characteristics (e.g., higher affinity maturation or fewer liabilities) (Figure 6) This process expands the candidate pool and aids the identification of candidates with improved characteristics. Likewise, single B cell and display methods also produce a shortlist of candidates that are experimentally tested and could be used to search back in the wider immune repertoire. Thus, efficient hit expansion with NGS repertoire data if performed effectively allows selecting antibodies with better properties. HIT EXPANSION Simply select better antibodies with the IGX Platform 21 Figure 6. The Phylogenetic Tree viewer in IGX-Branch enables users to quickly identify top antibody candidates based on the lineage and the visualization of antibody properties (including experimental metadata). The picture above shows an example of visualizing developability liabilities predicted with IGX-Annotate. Simply select better antibodies with the IGX Platform 22 Display technologies start from a synthetic, naive, or tailored library of antibody candidates which are screened against one or more target antigens. The method relies on (repeatedly) enriching the fraction of candidates that bind to the target of interest. However, relying solely on in vitro selection can yield a limited set of candidates, since the odds favor picking only the most highly represented candidates. As enrichment is affected by, among others, the initial clonal frequency, this can lead to promising candidates being missed. High “on-target” enrichment is not the only characteristic that defines a good therapeutic antibody, as spurious binding of irrelevant antigens is a frequent occurrence. Moreover, there might be a need to avoid binding of specific epitopes of the target antigen. By including additional panning conditions on “counter-targets” as well as “irrelevant” antigens, display technologies offer a degree of control far beyond what is available in immunized animals. NGS sequencing of the output of multiple rounds of various panning conditions (e.g., on-target, counter-target, irrelevant) increases the sensitivity and precision with which one can pick the best candidates. With increasing panning conditions and rounds, come additional datasets, which, in turn, increases the complexity of the data analysis and computational burden. Setting up an analysis including all panning conditions and rounds is technically challenging, due to the number of datasets, their scale, and the operations one must perform on them (i.e., join, intersect, subtract). Finally, incorporating knowledge of previously discovered candidates, wet-lab assay data, and structural predictions is crucial to making fully informed decisions and obtaining a high-quality lead list. The IGX Platform enables you to select the most promising antibody candidates from your display panning experiment by: ✓ Tracking clones across panning rounds to identify highly enriched candidates through interactive, user-defined visualizations ✓ Allowing users to freely join, intersect, and subtract your datasets to match your panning campaign design (Figure 7) ✓ Avoiding picking sub-optimal candidates by overlaying accurate liability predictions, assay data, and more onto enrichment graphs Together with the IGX Platform, NGS sequencing allows researchers to fully capitalize on the library diversity provided by display technology. It enables rational selection of enriched clones with increased diversity and quality of selected leads. ENRICHMENT ANALYSIS Simply select better antibodies with the IGX Platform 23 Figure 7. The Workflow Canvas in IGX-Track enables users to fully customize the data analysis to fit any panning campaign approach, no matter the complexity. Simply select better antibodies with the IGX Platform 24 ALL YOUR DATA IN ONE PLACE Antibody discovery workflows often utilize multiple sequencing technologies in parallel. As such, a complete data management and analysis tool needs to support many data types and formats. The IGX Platform supports data from a variety of sequencing technologies, including Sanger, bulk NGS, and single-cell sequencing (e.g., plate-based NGS or 10x Genomics). Data is harmonized upon import, allowing comparisons between datasets, and ensuring compatibility with all IGX Apps, removing barriers, and enabling scientists to fully utilize the strengths of different technologies. A great example is the common use case of ‘hit expansion’: small, well-characterized datasets, from a Sanger-based hybridoma approach or single-cell screening platform, can be used to ‘mine’ large-scale bulk NGS repertoires and find clonal relatives with better properties (e.g., better binding or fewer liabilities). This is done by clustering Sanger, single-cell NGS, and bulk NGS sequences together, a very laborious and complicated task when using publicly available tools or ad hoc solutions. The IGX Platform stores all your data in one place, and provides a powerful search engine, optimized for repertoire data. Forget looking for and searching through countless folders and Excel tables. Once you have analyzed the data with our specialized Apps, you can easily compare all potential candidates of interest to select the top leads for further development. In addition, IGX Platform facilitates data integration by providing easy-to-use import and export capabilities, including interactive metadata importers (see below). External Amazon S3 storage spaces can be easily linked to the platform, minimizing manual interactions and data transfers, and bringing increased productivity, security, and reducing chance of human error. MANAGE Simply select better antibodies with the IGX Platform 25 ✓ Synchronize with other systems: Maintain a single source of truth for all metainformation by automatically synchronizing clone-related data, including assay results and liability predictions, with systems like LIMS or ELN. Additionally, you can connect external data storages, such as Amazon S3, to automatically import or export your sequencing data. ✓ Automate: Streamline your analysis process by creating automated pipelines with an intuitive interface, while keeping the flexibility in exploring results and conducting additional case-specific analyses. ✓ Leverage custom code: Run proprietary algorithms to predict antibody properties, making its output available to scientists during candidate selection, and craft custom visualizations or analyses with the data. Alternatively, you can utilize a customized algorithm for sequence clustering and visualize its results in the IGX Platform interface. ✓ Data organization and Machine Learning: Leverage the scalable architecture and accessible interface of the IGX Platform with the powerful API and SDK to easily retrieve structured and labeled datasets to train your in-house Machine Learning models. This eliminates the time-consuming effort of structuring datasets and makes it simple to query your dataset for training. API FOR MAXIMUM PRODUCTIVITY Simply select better antibodies with the IGX Platform 26 METADATA AT ANY LEVEL Samples come with a plethora of associated metadata, such as assay results, clinical information, and in silico predicted properties. Only by closely associating this metadata to the sequencing data, a comprehensive analysis of the repertoire can be made. This integration is most crucial in antibody discovery workflows that need to condense the vast repertoire diversity into a handful of highly promising candidates. The IGX Platform provides an extensive tagging system that enables easy, flexible, and meaningful organization of antibody data and its associated metadata based on Tags. These are tightly integrated into the user experience and offer an intuitive way to add metadata to samples, cells, and sequences. The following three levels allow the user to add information relative to each of the functional levels of the antibody (Figure 8): ✓ Sample: any information relevant to all the sequences contained in a group of data (e.g., animal, patient, immunization timepoint, antigen, tissue) ✓ Clone: any information relevant to the full receptor formed by the combination of two chains (e.g., affinity, neutralization, expression, cluster Id) ✓ Sequence: any information relevant to the individual chain sequence (e.g., V call, J call, CDR3 amino acids, sequence count, liabilities presence Tags can not only be used to organize and retrieve data but also directly in visualizations as overlays or to stratify plots. This facilitates swift selection of clusters that contain the right samples or properties, and antibody clones with the maximum potential for clinical development. Adding tags is easy: you can manually add them to samples, clones, or sequences (one-by-one or in batch), or upload tables containing all metadata you would like to add. INTEGRATE Simply select better antibodies with the IGX Platform 27 Clinical information Experimental information Sequence information V Gene TRAV7-2 Tag Value Sequence level Affinity 230 Expression 1.72 Clone level Sample 1 Age 32 Tissue Lung Tag Value Sample level CARGNSGSYNRN Beta CQAWDSSTVVF Alpha CARGNSGSYNRN Beta CQUWDSSTVVF Alpha Figure 8. Adding metadata at different levels. Tags are tightly integrated into the user experience and offer an intuitive way to store metadata and create rich, interactive visualizations that use all available data. Simply select better antibodies with the IGX Platform 28 “By leveraging ENPICOM’s powerful data integration and analysis features, we will be able to increase our ability to unlock access to needle-in-a-haystack targets and continue to deliver the world’s most advanced antibody discovery solutions to our clients.” - Brian Walters, Chief Executive Officer at Genovac Simply select better antibodies with the IGX Platform 29 FIND WHAT YOU ARE LOOKING FOR With sequencing costs lower than ever, vast amounts of data can be generated. Antibody discovery workflows have been incorporating these large datasets into the selection and refinement workflows. With such increasing data quantities, being able to quickly find and select relevant samples for analysis is increasingly important. At the heart of the IGX Platform, the Clone Browser is a powerful tool to explore antibody sequencing data and its associated metadata. It is designed from the ground up to provide a structured overview of all data, while at the same time enabling extremely granular searches. You can easily find all samples in a project, find specific sequences of interest (e.g., all with a specific V-gene), or filter out sequences with undesirable characteristics (e.g., containing exposed liabilities). Single-cell and single-chain variable fragment (scFv) sequencing data are a challenge to manage and analyze, given the additional paired-chain information. The Clone Browser takes away the complexity of managing paired-chain sequencing data by automatically keeping track of the chain link. You can use this information throughout your analyses, for example when defining your clusters, or predicting exposed liabilities. COLLECT LEAD CANDIDATES INTO BASKET The Basket environment in the IGX Platform is like an advanced shopping cart for antibody candidates. It is the place dedicated to making the final selection of your candidates. Throughout the IGX Platform, interesting antibody candidates can be sent to Basket where an intuitive interface allows head-to-head comparison of the selected candidates. The page provides a customizable view of different types of data such as sequence (region) alignments, functional assay data obtained in the lab, or in silico predictions of exposed liabilities and developability metrics. Users can rank the listed antibodies based on the properties they value most and easily identify top candidates for further development. Simply select better antibodies with the IGX Platform 30 SHARE WITH COLLABORATORS AND CLIENTS For service providers, an intuitive and userfriendly environment to share data and showcase campaign results to clients significantly elevates the quality of service. Similarly, collaborations might require sharing specific datasets without compromising other data. Managing data transfer, access, and security is crucial, given its confidential nature. To enable easy and safe data collaboration and sharing, the IGX Platform introduces the concept of Data Spaces: virtual environments to store data and associated metadata in dedicated, secure locations. Data Spaces can be private to a single user, shared between project members, or used to share data with clients. Users can have access to one or more Data Spaces, meaning users can work on multiple projects at the same time, and restrict data to be shared externally with clients or collaborators in confined spaces. If required, Spaces can be made immutable, allowing sharing of data and antibody candidates, without the risk of modification by external parties. Simply select better antibodies with the IGX Platform 31 SECURE The IGX Platform is deployed using a Software as a Service (SaaS) model, hosted centrally on servers in Western Europe. This ensures that the IGX Platform can be accessed instantly and securely via a web browser. All authentication is securely handled by Auth0, providing enterpriselevel security with features such as two-factor authentication and single-sign-on. All data within the platform is shielded at three different levels: 1. Inter-service communication is completely contained within the platform and does not communicate with external networks 2. Inter-service requests are augmented with special session headers to ensure that only the intended recipients can access them 3. All data is encrypted at rest The cloud architecture ensures users always access the latest version of the platform and eliminates the need for on-site cluster infrastructure. This eliminates the need for continuous maintenance and upgrades to support the storage and analysis of evergrowing datasets. ENPICOM has a world-class security system in place and is ISO 27001 certified. ISO 27001 is an international standard on how to manage an ‘Information security management system’ (ISMS). ISMS defines and manages controls that an organization needs to implement to ensure that it is sensibly protecting the confidentiality, availability, and integrity of assets from threats and vulnerabilities. The ISO 27001 certification provides both new and existing clients with confidence and assurance in ENPICOM’s ability to protect their information, with demonstrable commitment to information security and privacy. As a client you want to know that your information is safe, this certification demonstrates ENPICOM’s commitment to having the highest standards in safeguarding our client information. AND ALWAYS ON Simply select better antibodies with the IGX Platform 32 TECHNICAL ASSISTANCE The IGX Platform and its Apps go through a thorough discovery process and extensive user testing to ensure a flexible and userfriendly experience. On top of that, we provide comprehensive customer support in the form of: Onboarding We provide online onboarding workshops of the IGX Platform and its applications which guide the user through curated use cases. Training We organize additional training sessions after the onboarding for specific Apps, workflows, or user groups to ensure users work efficiently. Intercom Instantly connect to a support operator directly from within the platform for any questions you might have. Knowledge base The IGX Platform includes a knowledge base with articles that provide in-depth descriptions of all the features available on the platform and explanations and videos on how to use different Apps. E-mail Our experienced customer support team is always available by email to provide tailored answers to complex questions. THAT GOES THE EXTRA MILE Simply select better antibodies with the IGX Platform 33 SUMMARY The IGX Platform and its Antibody Discovery Module are an end-to-end cloud solution that removes the complexity of managing and analyzing immune repertoire sequencing data for antibody discovery. It supports Sanger, single-cell, and NGS data, and can associate experimental results and clinical variables to the sequence data. By integrating all information in a single environment, the IGX Platform enables scientists to analyze their Rep-Seq data from all angles. Securely operating in the cloud, specialized Apps provide intuitive access to extensive analysis and visualization options. The IGX Platform gives every scientist the independence to simply select better antibodies. Set up, select, and process any data Any amplicon configuration can be set up, including UMIs, barcodes and linkers, so your analyses are performed seamlessly and accurately. Boost your candidate pool Let characterized Sanger hits guide the exploration of your large NGS outputs to expand and improve your candidate pool. Get the 360 view of your data Link assay and in silico data directly to clones, explore multiple antibody characteristics in a single visualization, and utilize an integrated scoring system to prioritize your top candidates. Identify specific and enriched candidates Seamlessly filter out candidates that bind undesired targets and define filter and sorting strategies to find the most enriched candidates. Interactive visualizations with rich metadata overlays help evaluate many properties at once and ease candidate selection. Simplify complex analyses Freely join, intersect, and subtract datasets, cluster clones together and plot phylogenetic trees to select better antibody candidates. Speed up candidate selection Use metadata annotations throughout the entire platform to prioritize clusters and sequences of interest, significantly reducing time spent on irrelevant data. De-risk antibody development Use high-throughput structural modeling to ensure accurate liability annotations and select diverse candidates with the best developability profiles. Skip the code Work with an intuitive user interface and accomplish complex tasks without translating biology to code, so you can focus on what matters most: your research. Stay in sync The IGX Platform runs securely in the cloud, which means you are always in sync and using the latest and greatest. Collaborate and share Communicate your insights with customers and other stakeholders through a lean, intuitive interface where they can explore their data with minimal training. Simply select better antibodies with the IGX Platform 34 REFERENCES 1. Kaplon H, Reichert JM. Antibodies to watch in 2019. MAbs [Internet]. 2019 Feb 17 [cited 2020 Dec 27];11(2):219–38. Available from: / pmc/articles/PMC6380461/?report=abstract 2. BIO, Biomedtracker, Amplion. Clinical Development Success Rates 2006-2015 - Pharma intelligence. 2016. 3. Bailly M, Mieczkowski C, Juan V, Metwally E, Tomazela D, Baker J, et al. Predicting Antibody Developability Profiles Through Early Stage Discovery Screening. MAbs [Internet]. 2020 Jan 1 [cited 2021 Jan 11];12(1):1743053. Available from: https:// www.tandfonline.com/doi/full/10.1080/19420 862.2020.1743053 4. Mackenzie M, Cambrosio A, Keating P. The commercial application of a scientific discovery: The case of the hybridoma technique. Res Policy. 1988 Jun 1;17(3):155–70. 5. Juy D, Legrain P, Cazenave PA, Buttin G. A new rapid rosette-forming cell micromethod for the detection of antibody-synthesizing hybridomas. J Immunol Methods. 1979 Nov 15;30(3):269–75. 6. Mcdaniel JR, DeKosky BJ, Tanno H, Ellington AD, Georgiou G. Ultra-high-throughput sequencing of the immune receptor repertoire from millions of lymphocytes. Nat Protoc. 2016 Mar 1;11(3):429–42. 7. McCafferty J, Griffiths AD, Winter G, Chiswell DJ. Phage antibodies: filamentous phage displaying antibody variable domains. Nature [Internet]. 1990 [cited 2020 Dec 27];348(6301):552–4. Available from: https:// pubmed.ncbi.nlm.nih.gov/2247164/ 8. Galson JD, Trück J, Clutterbuck EA, Fowler A, Cerundolo V, Pollard AJ, et al. B-cell repertoire dynamics after sequential hepatitis B vaccination and evidence for cross-reactive B-cell activation. Genome Med [Internet]. 2016 Jun 16 [cited 2021 Jan 14];8(1):68. Available from: http://genomemedicine. biomedcentral.com/articles/10.1186/s13073- 016-0322-z 9. Galson JD, Trück J, Clutterbuck EA, Fowler A, Cerundolo V, Pollard AJ, Lunter G, Kelly DF. B-cell repertoire dynamics after sequential hepatitis B vaccination and evidence for cross-reactive B-cell activation. Genome Med. 2016 Jun 16;8(1):68. doi: 10.1186/ s13073-016-0322-z. Erratum in: Genome Med. 2016;8(1):81. PMID: 27312086; PMCID: PMC4910312. 10. Phad GE, Pushparaj P, Tran K, Dubrovskaya V, Àdori M, Martinez-Murillo P, Vázquez Bernat N, Singh S, Dionne G, O’Dell S, Bhullar K, Narang S, Sorini C, Villablanca EJ, Sundling C, Murrell B, Mascola JR, Shapiro L, Pancera M, Martin M, Corcoran M, Wyatt RT, Karlsson Hedestam GB. Extensive dissemination and intraclonal maturation of HIV Env vaccineinduced B cell responses. J Exp Med. 2020 Feb 3;217(2):e20191155. doi: 10.1084/ jem.20191155. PMID: 31704807; PMCID: PMC7041718. igx@enpicom.com | +31852500575 | www.enpicom.com Sint Janssingel 88 | 5211 DA ‘s-Hertogenbosch | The Netherlands v09.23 SCAN HERE to download a digital copy