• Cagatay Turkay, University of Warwick, United Kingdom

• Hendrik Strobelt

• Michael Heinzinger, (TUM) Technical University of Munich, Germany
• Burkhard Rost, Rostlab, Germany
• Christian Dallago, (TUM) Technical University of Munich, Germany
• Tobias Olenyi, (TUM) Technical University of Munich, Germany
• Ahmed Elnaggar, (TUM) Technical University of Munich, Germany

With high throughput sequencing, quick insight into data separability for custom sequence datasets is desired to focus experiments on promising candidates. Recently, language models (LMs) have been adapted from use in natural language to work with protein sequences instead. Protein LMs show enormous potential in generating descriptive features for proteins from just their sequences at a fraction of the time of previous approaches. Protein LMs offer to convert amino acid sequences into embeddings that can also be used in combination with dimensionality reduction techniques (e.g. UMAP) to quickly span and visualize protein spaces (e.g. via scatter plots). On 3D scatter plots, proteins can be annotated with known properties to visually gain an intuition, even prior to training supervised models, about the separability of data. Additionally, conclusions can be drawn about proteins without annotation by putting them into the context of labelled proteins. The bio_embeddings pipeline offers an interface to simply and quickly embed large protein sets using protein LMs, to project the embeddings in lower dimensional spaces, and to visualize proteins in these spaces on interactive 3D scatter plots. The pipeline is accompanied by a web server that offers to visualize small protein datasets without the need to install software.

• Jon Hill, Boehringer Ingelheim Pharmaceuticals Inc, United States
• Di Feng, Boehringer Ingelheim Pharmaceuticals Inc, United States

Single-cell transcriptomics has become an increasingly common technique for understanding complex biological systems. Due to the high-dimensional data that is generated by such experiments, algorithms for dimensionality reduction, such as principal component analysis (PCA), t-distributed stochastic neighbor embedding (t-SNE), and uniform manifold approximation and project (UMAP), attempt to overcome these problems through projection of the high-dimensional findings into a lower-dimensional space.

As the output of these methods are commonly presented as scatterplots, they often suffer from overplotting, making nearby points difficult to distinguish. Since the spacing between points can be non-uniform, estimation of the relative proportions of cells in each cluster is not practical. Lack of consistent boundaries between points can also make it challenging to interpret cluster membership along with additional parameters, such as gene expression, in the same plot.

Here we present a complementary method to these clustering techniques for single cell data. By using linear assignment to map these projections to an appropriately-sized grid, it becomes possible to preserve overall cell-cell relationships in space, while condensing the space used for the figure, avoiding overplotting, and allowing for easy boundary annotation that permits overlay of additional data

• Inanc Birol, Canada's Michael Smith Genome Sciences Centre, Canada
• Rene L. Warren, BC Cancer Genome Sciences Centre., Canada
• Ka Ming Nip, Canada's Michael Smith Genome Sciences Centre, Canada
• Maria Stephenson, Canada's Michael Smith Genome Sciences Centre, Canada
• Saber Hafezqorani, Canada's Michael Smith Genome Sciences Centre, Canada
• Chen Yang, Canada's Michael Smith Genome Sciences Centre, Canada

Isoform detection and discovery at the single cell resolution are central to improving our understanding of heterogeneity in organs and tissues, and visualization tools would be instrumental in exploring this heterogeneity. However, current interactive transcriptome visualization tools are designed for bulk RNA-Seq data, and have limited utility in analyzing single-cell RNA-seq data. Here, we introduce RNA-Scoop, a visualization tool for single cell transcriptomics.

The input of RNA-Scoop is a single JSON file, which specifies the paths to a GTF file containing the isoforms of interest, a matrix file containing their expression levels in each cell, and files containing labels for the matrix rows and columns. Users can select genes for the isoform view, where all isoforms of selected genes are displayed. A t-SNE plot allows users to zoom in and out of different areas and select cells via lasso selection. Upon selection, displayed isoforms are colored according to their average level of expression in the selected cells. Expression per cluster is visualized through a dot plot. Additionally, isoforms are selectable, enabling users to highlight the cells in which isoforms of interest are expressed. Through these easy-to-use features, RNA-Scoop simplifies the interrogation of isoforms and cell types in thousands of cells.

• Antonios Somarakis, LUMC, Netherlands
• Vincent van Unen, LUMC, Netherlands
• Frits Koning, LUMC, Netherlands
• Boudewijn Lelieveldt, LUMC, Netherlands
• Thomas Höllt, TU Delft, LUMC, Netherlands

Tissue functionality is determined by the characteristics of tissue-resident cells and their interactions within their microenvironment. Imaging Mass Cytometry offers the opportunity to distinguish cell types with high precision and link them to their spatial location in intact tissues at sub-cellular resolution. This technology produces large amounts of spatially-resolved high-dimensional data, which constitutes a serious challenge for the data analysis. We present an interactive visual analysis workflow for the end-to-end analysis of Imaging Mass Cytometry data that was developed in close collaboration with domain expert partners. We implemented the presented workflow in an interactive visual analysis tool; ImaCytE. Our workflow is designed to allow the user to discriminate cell types according to their protein expression profiles and analyze their cellular microenvironments, aiding in the formulation or verification of hypotheses on tissue architecture and function. Finally, we show the effectiveness of our workflow and ImaCytE through a case study performed by a collaborating specialist. ImaCytE is open source and the code and binaries are available at https://github.com/biovault/ImaCytE.

• Marinka Zitnik

• Marwan Abdellah, Blue Brain Project / EPFL, Switzerland
• Nadir Román Guerrero, Blue Brain Project / EPFL, Switzerland
• Samule Lapere, Blue Brain Project / EPFL, Switzerland
• Jay S. Coggan, Blue Brain Project / EPFL, Switzerland
• Benoit Coste, Blue Brain Project / EPFL, Switzerland
• Snigdha Dagaer, Blue Brain Project / EPFL, Switzerland
• Daniel Keller, Blue Brain Project / EPFL, Switzerland
• Jean-Denis Courcol, Blue Brain Project / EPFL, Switzerland
• Henry Markram, Blue Brain Project / EPFL, Switzerland
• Felix Schurmann, Blue Brain Project / EPFL, Switzerland

Motivation: Accurate morphological models of brain vasculature are key to modeling and simulating cerebral blood flow (CBF) in realistic vascular networks. This in silico approach is fundamental to revealing the principles of neurovascular coupling (NVC). Validating those vascular morphologies entails performing certain visual analysis tasks that cannot be accomplished with generic visualization frameworks. This limitation has a substantial impact on the accuracy of the vascular models employed in the simulation. Results: We present VessMorphoVis, an integrated suite of toolboxes for interactive visualization and analysis of vast brain vascular networks represented by morphological graphs segmented originally from imaging or microscopy stacks. Our workflow leverages the outstanding potentials of Blender, aiming to establish an integrated, extensible and domain-specific framework capable of interactive visualization, analysis, repair, high-fidelity meshing and high-quality rendering of vascular morphologies. Based on the initial feedback of the users, we anticipate that our framework will be an essential component in vascular modeling and simulation in the future, filling a gap that is at present largely unfulfilled.
Availability and implementation: VessMorphoVis is freely available under the GNU public license on Github at https://github.com/BlueBrain/VessMorphoVis. The morphology analysis, visualization, meshing and rendering modules are implemented as an add-on for Blender 2.8 based on its Python API (Application Programming Interface). The add-on functionality is made available to users through an intuitive graphical user interface (GUI), as well as through exhaustive configuration files calling the API via a feature-rich CLI (command line interface) running Blender in background mode.

• Ivan Viola, King Abdullah University of Science and Technology, Saudi Arabia
• Tobias Klein, TU Wien, Austria
• M. Eduard Gröller, TU Wien, Austria
• Peter Mindek, TU Wien, Austria

Biologists often use computer graphics to visualize structures, which due to physical limitations are not possible to imagewith a microscope. One example for such structures are microtubules, which are present in every eukaryotic cell. They are part ofthe cytoskeleton maintaining the shape of the cell and playing a key role in the cell division. In this paper, we propose a scientifically-accurate multi-scale procedural model of microtubule dynamics as a novel application scenario for procedural animation, which cangenerate visualizations of their overall shape, molecular structure, as well as animations of the dynamic behaviour of their growth anddisassembly. The model is spanning from tens of micrometers down to atomic resolution. All the aspects of the model are driven byscientific data. The advantage over a traditional, manual animation approach is that when the underlying data change, for instance dueto new evidence, the model can be recreated immediately. The procedural animation concept is presented in its generic form, withseveral novel extensions, facilitating an easy translation to other domains with emergent multi-scale behavior.

• Dimitar Garkov, University of Konstanz, Germany
• Etienne Lein, Max Planck Institute of Animal Behaviour, Germany
• Christian Dullin, Medical Center Göttingen, Germany
• Karsten Klein, University of Konstanz, Germany
• Alex Jordan, Max Planck Institute of Animal Behaviour, Germany
• Falk Schreiber, University of Konstanz, Germany
• Bjorn Sommer, Royal College of Art, United Kingdom

Here we discuss our first prototype of an application for the analysis and visualisation of structural brain image data. The application is part of a processing pipeline to help researchers gain new insights by turning raw anatomical brain data into quantitative 3D representations. The pipeline has been optimised to process anatomical fish brain data obtained at a synchrotron imaging facility. The data is phase-retrieved, reconstructed and segmented before a three-dimensional mesh model is created. After being imported into the application, a model can be visualised in an immersive Virtual Reality environment on the HTC Vive headset. Each model can then be analysed by conducting a series of calibrated distance and volume measurements. The application is accompanied by an ImageJ plug-in, which supports users with image segmentation and model pre-processing.

• Mohammed El-Kebir, University of Illinois at Urbana Champaign, United States
• Jiaqi Wu, University of Illinois at Urbana Champaign, United States

Motivation: Cancer is caused by the accumulation of somatic mutations that lead to the formation of distinct populations of cells, called clones. The resulting clonal architecture is the main cause of relapse and resistance to treatment. With decreasing costs in DNA sequencing technology, rich cancer genomics datasets with many spatial sequencing samples are becoming increasingly available, enabling the inference of high-resolution tumor clones and prevalences across different spatial coordinates. While temporal and phylogenetic aspects of tumor evolution, such as clonal evolution over time and clonal response to treatment, are commonly visualized in various clonal evolution diagrams, visual analytics methods that reveal the spatial clonal architecture are missing.

Results: This paper introduces ClonArch, a web-based tool to interactively visualize the phylogenetic tree and spatial distribution of clones in a single tumor mass. ClonArch uses the marching squares algorithm to draw closed boundaries representing the presence of clones in a real or simulated tumor. ClonArch enables researchers to examine the spatial clonal architecture of a subset of relevant mutations at different prevalence thresholds and across multiple phylogenetic trees. In addition to simulated tumors with varying number of biopsies, we demonstrate the use of ClonArch on a hepatocellular carcinoma tumor with 280 sequencing biopsies. ClonArch provides an automated way to interactively examine the spatial clonal architecture of a tumor, facilitating clinical and biological interpretations of the spatial aspects of intra-tumor heterogeneity.

Availability: https://github.com/elkebir-group/ClonArch

• Florian Ganglberger, VRVis Research Center, Austria
• Joanna Kaczanowska, Research Institute of Molecular Pathology (IMP), Austria
• Wulf Haubensak, Research Institute of Molecular Pathology (IMP), Austria
• Katja Bühler, VRVis Research Center, Austria

Exploring the relationship between genes, brain circuitry, and behavior is a key element in neuroscience research. This requires joint analysis of heterogeneous spatial brain data, including 3D imaging data, anatomical data, and brain networks at varying scales, resolutions, and modalities. Current analytical workflows in neuroscience involve time-consuming manual aggregation of the data and only sparsely incorporate spatial context to operate continuously on multiple scales. Incorporating techniques for handling spatial brain data is therefore a necessity.
We present a novel web-based framework to explore heterogeneous neurobiological spatial data of different types, modalities and scale for interactive visual analytics workflows. It enables domain experts to combine data from large-scale brain initiative by utilizing the hierarchical and spatial organization of the data. Connectivity data at different resolutions, such as mesoscale structural connectivity and region-wise functional connectivity can be accessed on a common hierarchical reference space. On-demand queries on volumetric gene expression and connectivity data enable an interactive dissection of networks, with billions of edges, in real-time.
We demonstrate the relevance of our approach by reproducing findings of known microcircuits in fear and reward-system related functional neuroanatomy in mice. Further, we show its versatility by comparing multimodal brain networks linked to autism.

• Andrea Chatrian
• Jens Rittscher, University of Oxford, United Kingdom
• Heba Sailem, University of Oxford, United Kingdom

Intuitive visualisation of microscopy data is important for facilitating results interpretation, and identification of interesting patterns. Existing visualisation approaches, such as bar charts and heat maps, do not accommodate the complexity of visual information present in microscopy data. Previously we developed PhenoPlot; the first tool of its kind for visualising cellular imaging data. Here we develop PhenoPlot-v2 that allows visualisation of various structural entities at the molecular, cellular, and/or tissue level. PhenoPlot allows representing the hierarchy or interaction between various objects such as tissue, cells, and nuclei. The user can add multiple objects and define their shapes. For example, an ellipse can be used to represent a cell and a rectangle can represent a tube. If required, the user can draw their object for more complex shapes such as neurons. Multiple visual elements can be defined for each object including object dimensions, colours, pattern filling, line features, and extrusions. The user can map the measured image features to a selection of visual elements resulting in a pictorial representation of these objects. We demonstrate that PhenoPlot aids the identification of interesting patterns when applied to datasets describing phenotypes at the cellular and tissue level.

• Kari Lavikka, University of Helsinki, Finland
• Jaana Oikkonen, University of Helsinki, Finland
• Rainer Lehtonen, University of Helsinki, Finland
• Johanna Hynninen, Turku University Hospital, Finland
• Sakari Hietanen, Turku University Hospital, Finland
• Sampsa Hautaniemi, University of Helsinki, Finland

Visualization is the primary method in the exploration of genomic aberrations of cancer samples. However, the current state-of-the-art interactive genomic visualization tools (specifically genome browsers) have been designed for specific data formats, which are displayed in very definite, often rigid ways. Adapting the tools for atypical data and rich visual representations is difficult.

Here, we present GenomeSpy, a declarative, grammar-based approach for specifying interactive genomic visualizations that use the graphics processing unit (GPU) in rendering. With combinatorial building blocks such as graphical marks, scales, view composition operators, and transformations, the user can create novel, more appropriate visualizations. Fluid interactions, along with high rendering performance, help the user stay in the flow of exploration.

We demonstrate GenomeSpy with two case studies involving high-grade serous ovarian cancer data. First, we used GenomeSpy to create a tool for scrutinizing raw copy-number variation (CNV) data along with segmentation results. Second, we used the CNV results and point mutations in a multi-sample visualization that allows for exploring and comparing both multiple data dimensions and samples at the same time.

Although our focus has been on cancer research, we envision GenomeSpy being applied to other domains as well. The software is available at https://genomespy.app/.

• Hsiang-Yun Wu, TU Wien, Austria
• Martin Nöllenburg, TU Wien, Austria
• Filipa L. Sousa, University of Vienna, Austria
• Ivan Viola, King Abdullah University of Science and Technology, Saudi Arabia

Biological pathways represent chains of molecular interactions in biological systems that jointly form complex dynamic networks. The network structure changes from the significance of biological experiments and layout algorithms often sacrifice low-level details to maintain high-level information, which complicates the entire image to large biochemical systems such as human metabolic pathways. Our work is inspired by concepts from urban planning since we create a visual hierarchy of biological pathways, which is analogous to city blocks and grid-like road networks in an urban area. We automatize the manual drawing process of biologists by first partitioning the map domain into multiple sub-blocks, and then building the corresponding pathways by routing edges schematically, to maintain the global and local context simultaneously. Our system incorporates constrained floor-planning and network-flow algorithms to optimize the layout of sub-blocks and to distribute the edge density along the map domain. We have developed the approach in close collaboration with domain experts and present their feedback on the pathway diagrams based on selected use cases. We present a new approach for computing biological pathway maps that untangles visual clutter by decomposing large networks into semantic sub-networks and bundling long edges to create space for presenting relationships systematically.

• Andrea Guarracino, Centre for Molecular Bioinformatics, University Of Rome Tor Vergata, Rome, Italy, Italy
• Simon Heumos, University of Tubingen, United States
• Josiah Seaman, Max Planck Institute, Department of Molecular Biology, United States
• Eric Garrison, University of California Santa Cruz, United States

Pantograph is an open source browsable pangenome visualization for graph genomes. It allows researchers to see the full genetic diversity in large populations. Graph genomes naturally express genome rearrangements, SNPs, and indels. Using a graph structure provides integration with knowledge graphs containing annotations, geographical locations, and patient outcomes. This makes Pantograph an ideal tool for tracking viral strains of SARS-CoV-2. Without the constraints of a reference genome, viral strains can be smoothly integrated as they are sequenced.

Pantograph is a data and visualization solution that scales to thousands of individuals while preserving all types of sequence variation. Other tools either are not scalable or discard all genome rearrangements. Pantograph achieves scalability by identifying syntenic blocks and interconnecting them with nonlinear variants.

Pantograph’s application to the COVID-19 pandemic is driven by the unique evolutionary scenario we are facing. The infected population size is a multiplier for the number of mutations available for selection. Vaccinating the population during a pandemic is a selection sweep for resistant viral mutations. This is the biggest viral selection sweep in human history. Pantograph can help predict vaccine effectiveness in different regions of the world by integrating our total knowledge of genetic diversity.

• Anthony Federico, Boston University, United States
• Stefano Monti, Boston University, United States

Network analysis is a growing trend in genomics whereby emphasis is placed on unraveling the relationships between genes and detecting changes in connectivity across phenotypes. It remains a challenge to effectively analyze genome-wide regulatory networks, which typically take the form of large and highly connected graph structures – which is further exacerbated when comparing multiple networks simultaneously. It is thus crucial to develop resources geared towards the efficient analyses and comparison of biological regulatory networks. To that end, we have developed Netviz, an interactive R Shiny web application for the analysis of regulatory networks. Netviz provides an intuitive interface for analyses guided by graph theoretic approaches and is paired with an efficient backend capable of handling tens of thousands of nodes. While there is a strong emphasis on gene- and protein- level networks, many methods are node-agnostic and compatible with any regulatory layer. Some features include methods for in-silico validation of predicted interactions, inference of unknown functional properties based on network localization, centrality-based node scoring, and interactive sub-network visualization and enrichment. Netviz is open-source, freely available as an R package, and supports all major platforms. It remains under active development in pace with the evolving field of network genomics.

• Selim Kalayci, Department of Genetics and Genomic Sciences, Icahn School of Medicine at Mount Sinai, New York, NY, 10029, USA, United States
• Francesca Petralia, Department of Genetics and Genomic Sciences, Icahn School of Medicine at Mount Sinai, New York, NY, 10029, USA, United States
• Pei Wang, Department of Genetics and Genomic Sciences, Icahn School of Medicine at Mount Sinai, New York, NY, 10029, USA, United States
• Zeynep Gümüş, Icahn School of Medicine at Mount Sinai, United States

To better understand the molecular basis of cancer, the NCI’s Clinical Proteomics Tumor Analysis Consortium (CPTAC) has been performing comprehensive large-scale proteogenomic characterizations of multiple cancer types. Gene and protein regulatory networks are subsequently being derived based on these proteogenomic profiles, which serve as tools to gain systems-level understanding of the molecular regulatory factories underlying these diseases. On the other hand, it remains a challenge to effectively visualize and navigate the resulting network models, which capture higher order structures in the proteogenomic profiles. There is a pressing need to have a new open community resource tool for intuitive visual exploration, interpretation and communication of these gene/protein regulatory networks by the cancer research community. In this work, we introduce ProNetView-ccRCC (http://ccrcc.cptac-network-view.org/), an interactive web-based network exploration portal for investigating phosphopeptide co-expression network inferred based on the CPTAC clear cell renal cell carcinoma (ccRCC) phosphoproteomics data. ProNetView-ccRCC enables quick, user-intuitive visual interactions with the ccRCC tumor phosphoprotein co-expression network comprised of 3,614 genes, as well as 30 functional pathway-enriched network modules. Users can interact with the network portal and can conveniently query for association between abundance of each phosphopeptide in the network and clinical variables such as tumor grade.

• Fritz Lekschas, Harvard University, United States

Epigenomic datasets contain rich sets of patterns that can act as proxies for non-coding regulatory elements. Visually searching for patterns by similarity can be challenging because of the large search space, the visual complexity of patterns, and the user's perception of similarity.
I will present Peax (http://peax.lekschas.de), a new tool for visual pattern search in sequential epigenomic datasets. Peax is based on a convolutional autoencoder for unsupervised representation learning of regions in the dataset that can capture more visual details of complex patterns compared to existing similarity measures. Using this learned representation as features of the data, our accompanying visual query system enables interactive feedback-driven adjustments of the pattern search to adapt to the users' perceived similarity. Peax collects user-generated binary relevance feedback for genomic regions sampled with active learning. Using the feedback, Peax trains a model for binary classification to ultimately find other regions that exhibit patterns similar to the search target. In two user studies, we found that Peax's learned feature representation retrieves significantly more similar patterns than other commonly-used techniques. In this talk, I will demonstrate Peax's features through examples with ChIP-seq and DNase-seq data and report our insights on a user study with eight computational biologists.

• Cagatay Turkay, University of Warwick, United Kingdom