# Ondrej Strnad > Research Scientist / Engineer / Full Stack Developer. I turn hard technical problems into software people use – from electron-microscope control systems to GPU-rendered virtual cells and AI-native products. Website: https://ondrejstrnad.com/ LinkedIn: https://www.linkedin.com/in/ondrejstrnad/ Facebook: https://www.facebook.com/ondrej.strnad.33 Instagram: https://www.instagram.com/ondrej__strnad/ GitHub: https://github.com/ondrejstrnad ORCID: https://orcid.org/0000-0002-8077-4692 ## Expertise - Desktop Applications Development: C#, C/C++, Qt, WinForms, WPF, WCF, Windows, Windows Service, OpenGL, Vulkan, GLSL, Java, NetBeans, OpenVR/OpenXR, HoloLens - Full Stack Web Development: React, TypeScript, JavaScript, HTML5, CSS3, SASS, Python, SQL, MongoDB, PHP, WebGPU, WebAssembly, Node.js, GraphQL - DevOps & Automation: Git, GitHub Actions, Linux, Render.com - Team Coordination & Delivery: Small team leadership, End-to-end feature delivery, Sprint/task planning, Stakeholder communication, Cross-team coordination, Hands-on mentoring, Release planning, Shared delivery ownership, Risk management, Process improvement ## Career history ### Senior Research Engineer, Mohamed bin Zayed University of Artificial Intelligence (MBZUAI), UAE (02/2026 – present) Link: https://mbzuai.ac.ae - Lead developer of CellMap – the WebGPU visualisation engine of the AIDO Virtual Cell Lab, taking AI predictions from whole cells down to atoms in the browser. - Main author of the rendering codebase; built it as a reusable library so product teams embed one engine across the platform. ### Research Scientist, King Abdullah University of Science and Technology (KAUST), KSA (04/2023 – 01/2026) Link: https://www.vccvisualization.org - Delivered TeraCellViewer, an interactive viewer for tera-scale oil reservoir simulations in an industry collaboration. - Turned research on neural flow representations into working, GPU-accelerated software. ### Research Scientist, King Abdullah University of Science and Technology (KAUST), KSA (11/2019 – 03/2023) Link: https://www.nanovis.org - Led development of MesoCraft – procedural modelling of billion-atom biological scenes, shipped on desktop and the web. - Software lead behind a series of IEEE TVCG publications, working hand-in-hand with PhD students to turn prototypes into usable tools. ### Visiting Researcher, King Abdullah University of Science and Technology (KAUST), KSA (02/2019 – 05/2019) Link: https://www.kaust.edu.sa ### Senior Software Developer, DELONG INSTRUMENTS a.s., Brno, Czech Republic (05/2014 – 10/2019) Link: https://www.delong.cz - Developed microCS, the control system shared across the company's electron-microscope product line. - Delivered internal systems that took manual work out of production – Softeee (release & configuration management), LabSoft (manufacturing records & warehouse), VGraph and LabCAM. ### Backend Developer, AdBros s.r.o., Brno, Czech Republic (06/2016 – 08/2018) Link: https://www.adbros.cz - Backend development for commercial web projects, in parallel with my role at DELONG. ### Researcher, Software Developer, co-founder, CaverSoft s.r.o., Brno, Czech Republic (08/2011 – 09/2015) Link: https://www.caver.cz - Co-founded the company commercialising the CAVER protein-analysis tools. - Built CAVER Analyst, the interactive desktop application that made tunnel analysis accessible to biochemists without programming. ### Software Developer, DELONG INSTRUMENTS a.s., Brno, Czech Republic (03/2007 – 05/2014) Link: https://www.delong.cz - Built ImPress, a visual programming environment that let test engineers create hardware test sequences without a developer. - Seven years building instrument control and test software, leading to the senior role. ## Projects ### Tandem The two-panel file manager for macOS that your AI can drive – while you stay in control. - Role: Founder – product, design and engineering - Context: Own product - Period: 2026 - Link: https://tandem-site.fly.dev - Technology: Swift, AppKit, MCP, Python plugins, Sparkle, Fly.io Challenge: People who work with files all day miss the keyboard-driven, two-panel managers they know from Windows, and Finder does not replace them. At the same time AI agents are starting to work with local files – usually from a terminal, where the user cannot see what is about to change. What I delivered: - A native macOS file manager in the Altap Salamander tradition: two panels, a function-key bar, keyboard profiles for Salamander, Total Commander and Mac habits, and seven themes. - The everyday tools built in – instant previews (a 10 GB log opens at once), archives browsed as folders, content search, batch rename and image tools, and an HTTP request window for developers. - An MCP server that lets Claude Code, Claude Desktop and Cursor see the panels and propose changes; every change opens Tandem's own dialog and waits for the user. - Guard rails for agents – a read-only mode, hidden folders and an activity log of every call – plus a module system with Python plugins. - The business around it: website, licensing with a 14-day trial and one-time pricing, and automatic updates. Business impact: - Makes AI file work visible and reversible: the agent fills in the dialog, the person presses Enter. - Replaces several single-purpose utilities – archiver, batch renamer, image converter, search tool – with one keyboard-driven app. - Private by design: no telemetry, no account, and the agent connection is a local socket only. ### Vistune A music and movie player for your cloud library – and, on top of it, a coaching tool for frame-by-frame review and side-by-side comparison. - Role: Founder – product, design and engineering - Context: Own product - Period: 2026 - Technology: TypeScript, React Native, Expo, Expo Router, AVFoundation, StoreKit 2, OneDrive & Google Drive APIs, EAS Challenge: Music, films and training footage tend to sit in the same cloud drive, yet each needs a different app. Coaches film their athletes all the time, but the footage ends up in a cloud folder and the tools to analyse it are either expensive desktop software or apps that want the videos uploaded to their servers. Showing an athlete one frame, one angle or two attempts next to each other should take seconds on a phone. What I delivered: - A music library player for the songs in OneDrive or Google Drive: background audio with lock-screen controls, queue, shuffle, repeat, playlists and offline downloads. - A movie library player for the films in the same drives: streaming or offline, landscape fullscreen, subtitles and speed control. - On top of the players, a coaching Studio that opens video straight from OneDrive, Google Drive, Photos or Files and steps through it frame by frame, with slow motion from 0.25× to 2×. - Side-by-side comparison of two videos with a locked offset – they play, pause and step in sync. - Drawing tools on the frame (lines, angles with degree readout, circles, arrows, freehand) and export of annotated frames and clips, rendered on the device. - A free app with a Pro subscription built on StoreKit 2 and verified on the device – no server, no analytics and no third-party purchase SDK. Business impact: - One app for three jobs: the music library, the movie library and coaching video analysis. - A coach can review a serve, a swing or a lift on the spot and send the annotated clip to the athlete from the same phone. - The athlete's footage stays where it is: cloud files are opened read-only and nothing is uploaded to the developer. - No backend to run – the whole product is the app, which keeps operating costs close to zero. ### Compeetee The operating system for running sports competitions and clubs. - Role: Founder – product owner and lead engineer - Context: Own product - Link: https://www.compeetee.com - Technology: TypeScript, React, React Native, Expo, Node.js, Express, GraphQL, MongoDB Atlas, Stripe, Fly.io, CI/CD Challenge: Sports clubs and leagues run competitions on spreadsheets, group chats and paper scorecards. Organisers lose hours to admin, players wait for results, and clubs have no data on their members' engagement or progress. What I delivered: - Designed and built the whole platform: web app, REST/GraphQL API and native iOS and Android apps, with payments, five UI languages and an AI coach. - End-to-end competition flow – registration, draws, schedules, live scoring, public leaderboards, seasons and rankings – plus club, training and analytics tools. - Engineering discipline of a much larger team: ~5,000 automated tests, CI/CD, staged releases and in-app tutorials for every major workflow. - Integrated my other products – OpenGC for course data, ResponseDock for club surveys – into one ecosystem. Business impact: - Used to run real golf events at Safaa Golf Club (KAUST), including the KAUST Cup and the Race to KAUST Cup 2025–26 season. - A competition can be set up in minutes and followed live by everyone on a single public link – no more chasing scorecards. - Clubs get member-engagement and performance data they never had before, turning admin into insight. ### OpenGC An open, community-driven database of the world's golf courses. - Role: Founder – product and engineering - Context: Own product - Link: https://www.opengc.net - Technology: TypeScript, React, Material UI, Node.js, MongoDB Atlas, REST API, AI / OCR, Leaflet, CI/CD Challenge: Detailed golf course data – hole layouts, distances, par, tees – is locked behind paid, closed databases. Every golf app has to buy or rebuild it, and golfers cannot contribute corrections. What I delivered: - An open database with a public API, SDK and embeddable widgets (maps, explorer, personal stats). - AI-assisted data entry: golfers upload a scorecard photo, OCR and AI extract the holes, and the system scores completeness and duplicate risk. - A moderation workflow (approve, request changes, merge duplicates) with versioned scorecards to keep the data trustworthy. Business impact: - Over 32,000 clubs listed – a data asset built by the community instead of bought from a vendor. - Became the course-data backbone of Compeetee: a course added in OpenGC appears in Compeetee automatically. - AI turns a photo into structured data in seconds, cutting the cost of each contribution to almost zero. ### ResponseDock A white-label survey engine that platforms embed in their own product – builder, embeddable forms and live results, with a browser SDK and React components. - Role: Founder – product and engineering - Context: Own product - Link: https://www.responsedock.com - Technology: TypeScript, React, Material UI, Node.js, Fastify, Prisma, MongoDB Atlas, OpenAI API, Stripe, SDK, React Native Challenge: Platforms and SaaS teams want their customers to run surveys inside the product, under the platform's own brand. Off-the-shelf survey tools send users to a third-party site with a second login, and building a survey builder in-house costs months of engineering. What I delivered: - A complete survey engine to embed: the survey builder for managers, respondent forms and live results with CSV, Excel and PDF export. - Embeddable forms – a survey, poll or registration form shown inline, in a modal or on a hosted page, themed before first paint and resized to its content, in 18 respondent languages including right-to-left. - Developer-first integration: a zero-dependency TypeScript browser SDK under 8 kB gzipped, loaded with one script tag, and React components (embed, modal, provider and hook), packaged for npm as @responsedock/sdk-browser and @responsedock/sdk-react. - Sign-in from the host app with short-lived signed sessions and per-user permissions, webhooks and signed callbacks, and Slack, Teams, Google Sheets, Jira and Linear integrations. - AI-assisted survey design and translation, so non-experts produce good surveys in any language. Business impact: - Lets a platform ship in-app surveys in an afternoon instead of building feedback tooling for months. - Respondents only ever see the host's brand and domain – no second login and no detour to a third-party site. - Dog-fooded in Compeetee for club surveys – one engine reused across products. ### TeraCellViewer Interactive exploration of tera-scale oil reservoir simulations. - Role: Research scientist – architecture and implementation - Context: KAUST Visual Computing Center · industry collaboration - Period: 2023 – 2026 - Technology: C++, Vulkan, Qt, GLSL, VR/XR Challenge: Oil reservoir simulations are too large to fit in memory, yet engineers need to inspect them interactively to make costly drilling and production decisions. What I delivered: - An out-of-core streaming renderer that keeps tera-scale models interactive on a workstation. - Multi-attribute views, split panels and integrated Vega-Lite charting in one analyst tool, with VR/XR support. Business impact: - Engineers explore full-resolution models directly, instead of waiting for down-sampled extracts. - Combines 3D views and charts in one place, shortening the path from raw simulation data to a decision. ### MesoCraft Procedural modelling of billion-atom biological scenes, on desktop and the web. - Role: Research scientist – lead developer - Context: KAUST Nanovisualization group - Period: 2019 – 2023 - Link: https://mesocraft.kaust.edu.sa - Technology: C++, OpenGL, Qt, GLSL, WebGPU, WebAssembly, VR/XR Challenge: Biologists want to see whole viruses and cells at atomic detail, but building such models by hand takes months and the scenes – billions of atoms – overwhelm standard tools. What I delivered: - Procedural generation that builds full biological assemblies from rules and data in minutes. - A rendering stack that keeps billion-atom scenes interactive – native (C++/OpenGL) and in the browser (WebGPU/WebAssembly). - The software foundation behind a series of peer-reviewed publications (IEEE TVCG and others). Business impact: - Turned month-long manual modelling into an interactive workflow researchers can run themselves. - Web deployment made the models accessible to collaborators and the public with nothing to install. ### DevelSlave A free developer workspace of 15 tool suites for the small tasks that interrupt real work – formatting, validation, conversion, inspection and generation. - Role: Author – product, design and engineering - Context: Own product - Link: https://www.develslave.com - Technology: TypeScript, React, Material UI, Node.js, CI/CD Challenge: Formatting JSON, decoding a JWT, hashing a string, converting a timestamp, generating test data – small tasks that pull developers out of flow many times a day, usually into a different ad-ridden site each time, with no idea where the pasted data ends up. What I delivered: - One workspace with 15 tool suites in four categories: text and generators, data and formats (JSON, XML, CSV/SQL, binary, units), security and quality (hashes, validation, colours) and web and network (URLs, headers, JWT, DNS, subnets). - Local-first by design: most tools run directly in the browser, with server calls only where the task needs them, such as DNS lookups. - Tools that connect – a value moves from one tool to the next (text to hash, JSON to text, URL to DNS lookup) without copying it by hand. - Free to use with no signup, search across all tools, and light and dark themes. Business impact: - One bookmark replaces a dozen single-purpose sites for everyday formatting, conversion and debugging. - Pasted data stays in the browser for most tools, which matters when the input is a token, a password or customer data. - The explicit hand-offs between tools make it usable for teaching, too: students see why each step exists. ### CAVER / CAVER Analyst Software for analysing tunnels and channels in proteins – used in drug design and enzyme engineering. - Role: Co-founder, researcher and lead developer of CAVER Analyst - Context: CaverSoft · Masaryk University - Period: 2011 – 2015 - Link: https://www.caver.cz - Technology: Java, NetBeans Platform, JOGL Challenge: How molecules travel through a protein decides how well a drug binds or an enzyme works – but tunnel analysis was slow, command-line driven and hard to interpret. What I delivered: - Co-founded CaverSoft to commercialise the CAVER tools. - Built CAVER Analyst: an interactive desktop application to compute, filter and explore protein tunnels in static structures and simulations. Business impact: - Made tunnel analysis accessible to biochemists without programming – part of a toolset used by research groups worldwide. - Results published in leading journals (PLoS Computational Biology, Bioinformatics, BMC Bioinformatics). ### microCS The control system behind DELONG's electron microscopes. - Role: Senior software developer - Context: DELONG INSTRUMENTS a.s. - Period: 2007 – 2019 - Technology: C/C#, .NET, WinForms, WCF, Windows Service, Protobuf Challenge: Electron microscopes combine precision hardware, cameras from several vendors and automated procedures that must run reliably under strict safety constraints. What I delivered: - A unified control system orchestrating hardware communication, procedure execution and the operator GUI across the product line. - Service-based architecture (Windows Service, WCF, Protobuf) separating the hardware layer from the user interface. Business impact: - One software platform for the whole product line instead of per-instrument code – faster to extend for new devices. - Reliable automated procedures reduced the operator skill needed to get good results. Proprietary tool of DELONG INSTRUMENTS a.s. ### Softeee Release and configuration management for manufactured instruments. - Role: Senior software developer - Context: DELONG INSTRUMENTS a.s. - Technology: C#, WPF, Entity Framework, MySQL Challenge: Shipping the wrong firmware or configuration to a precision instrument is expensive to fix in the field. What I delivered: - A system managing versioned releases of microCS and device-specific scripts for every hardware revision. Business impact: - Every instrument leaves the factory with a validated, traceable software configuration. - Support can see exactly what a customer's device runs – faster diagnosis, fewer site visits. Proprietary tool of DELONG INSTRUMENTS a.s. ### LabSoft Manufacturing tasks, warehouse and per-device production records in one system. - Role: Senior software developer - Context: DELONG INSTRUMENTS a.s. - Technology: C#, .NET, WinForms, WPF, MySQL Challenge: Production coordination relied on spreadsheets and paper checklists – error-prone and impossible to audit. What I delivered: - A central system for task management, warehouse inventory and a manufacturing diary for each device. Business impact: - Full audit trail of how each instrument was built. - Fewer coordination errors and less time spent tracking parts and status. Proprietary tool of DELONG INSTRUMENTS a.s. ### ImPress Visual programming for hardware test sequences. - Role: Software developer - Context: DELONG INSTRUMENTS a.s. - Technology: C#, .NET, WinForms Challenge: Every hardware test scenario needed a bespoke script written by an engineer – slow and error-prone. What I delivered: - A visual environment where test sequences are assembled graphically and executed against devices through microCS. Business impact: - Test engineers build and change tests themselves, without waiting for a developer. Proprietary tool of DELONG INSTRUMENTS a.s. ### VGraph Turning production logs into readable, zoomable graphs. - Role: Software developer - Context: DELONG INSTRUMENTS a.s. - Period: 2007 – 2019 - Technology: C#, .NET, WinForms Challenge: Manufacturing logs hide signals about quality and process drift, but raw logs are unreadable. What I delivered: - A log parser and graphing tool for production data. Business impact: - Engineers spot trends and anomalies during production instead of after a failure. Proprietary tool of DELONG INSTRUMENTS a.s. ### LabCAM One acquisition interface for cameras from many vendors. - Role: Software developer - Context: DELONG INSTRUMENTS a.s. - Period: 2007 – 2019 - Technology: C#, .NET, WinForms Challenge: Each camera vendor brings its own drivers and image formats, fragmenting the lab workflow. What I delivered: - A unified acquisition layer that abstracts vendor differences. Business impact: - Freedom to choose or switch camera vendors without rewriting the workflow; directly comparable results. Proprietary tool of DELONG INSTRUMENTS a.s. ## Education - PhD Computer Science, Faculty of Informatics, Masaryk University, Brno (2009 – 2014). Thesis: Algorithms for Detecting Pathways in Large Protein Structures and Their Ensembles (https://is.muni.cz/th/lkrh1/thesis_strnad.pdf?info) - MSc Computer Graphics, Faculty of Informatics, Masaryk University, Brno (2007 – 2009). Thesis: Free-form Deformations with Catmull-Clark Volumes (https://is.muni.cz/th/py7bf/diplomka.pdf?info) - BSc Applied Informatics, Faculty of Informatics, Masaryk University, Brno (2004 – 2007). Thesis: Sweep-based Freeform Deformations (https://is.muni.cz/th/139568/fi_b/bcprace.pdf?info) - Graduation – Sport education, theory of sport, Sport High School, Pardubice, Czech Republic (2004 – 2009) - Online Courses, ReactJS, TypeScript, CI/CD for Software Developers, Neural Networks and Deep Learning ## Publications ### 2026 #### CellMap: WebGPU Visualization of Multi-Scale AI-Generated Cell Biology Data Ondřej Strnad, Deng Luo, Venkatesh Nair, Le Song. Web3D '26: The 31st ACM International Conference on 3D Web Technology. DOI: https://doi.org/10.1145/3822516.3834870 Link: https://dl.acm.org/doi/full/10.1145/3822516.3834870 CellMap is a native browser visualization system for exploring AI-generated biological data across multiple spatial scales. It integrates outputs from protein localization predictors, abundance estimators, perturbation simulations, and cell morphology generators into a unified visual representation. The system converts heterogeneous AI-generated outputs into lightweight proxy geometry enriched with biological metadata, enabling efficient transmission, rendering, and interaction in the browser. A WebGPU renderer then provides scale-dependent visualization, supporting continuous navigation from cell populations to intracellular structures and localized proteins. ### 2025 #### AdaptiveFlowNet: A Hybrid Neural Vector Field Representation for Interactive Flow Visualization Amani Ageeli, Ondřej Peterka, Ondřej Strnad, Xingdi Zhang, Peter Rautek, Ali Awami, Florian Mannuss, Markus Hadwiger. Venue to be specified. [under review] Large-scale fluid simulations generate time-varying vector fields that contain highly dynamic and complex regions, which require substantial storage for accurate representation. Neural network models can provide compact representations of vector fields. Still, they often fail to capture complex flow behavior reliably, and their inference times can be too long for real-time visualization. This limitation is particularly evident in turbulent or rapidly changing regions. To address these challenges, we propose a hybrid neural vector field representation that adaptively combines a compact neural model with selective high-resolution storage of local Eulerian grids. The goal of the pipeline is to achieve a balance between accuracy and memory efficiency. Our method introduces a flow complexity mask that identifies regions where neural inference is unreliable. This mask can be derived from multiple sources, for example, by analyzing areas where the neural network shows high reconstruction error during training, from user-provided annotations, or from established flow metrics like the Finite-Time Lyapunov Exponent (FTLE). During the streamline reconstruction, regions marked as complex are reconstructed from local Euelrian grids, while the remaining areas are reconstructed using neural inference. In general, the method supports representations that vary in memory size, which provides a balance between storage requirements and fidelity. Furthermore, it supports dynamic memory allocation and GPU-based visualization, and it adapts to temporal changes in the vector field without the need to retrain the neural model. #### Nanouniverse: Virtual Instancing of Structural Detail and Adaptive Shell Mapping Ruwayda Alharbi, Ondřej Strnad, Markus Hadwiger, Ivan Viola. IEEE Transactions on Visualization and Computer Graphics. DOI: https://doi.org/10.1109/TVCG.2025.3618914 Rendering huge biological scenes with atomistic detail presents a significant challenge in molecular visualization due to the memory limitations inherent in traditional rendering approaches. In this paper, we propose a novel method for the interactive rendering of massive molecular scenes based on hardware-accelerated ray tracing. Our approach circumvents GPU memory constraints by introducing virtual instantiation of full-detail scene elements. Using instancing significantly reduces memory consumption while preserving the full atomistic detail of scenes comprising trillions of atoms, with interactive rendering performance and completely free user exploration. We utilize coarse meshes as proxy geometries to approximate the overall shape of biological compartments, and access all atomistic detail dynamically during ray tracing. We do this via a novel adaptive technique utilizing a volumetric shell layer of prisms extruded around proxy geometry triangles, and a virtual volume grid for the interior of each compartment. Our algorithm scales to enormous molecular scenes with minimal memory consumption and the potential to accommodate even larger scenes. Our method also supports advanced effects such as clipping planes and animations. We demonstrate the efficiency and scalability of our approach by rendering tens of instances of Red Blood Cell and SARS-CoV-2 models theoretically containing more than 20 trillion atoms. #### SynopFrame: Multiscale Time-dependent Visual Abstraction Framework for Analyzing DNA Nanotechnology Simulations. Deng Luo, Alexandre Kouyoumdjian, Ondřej Strnad, Haichao Miao, Ivan Barišić, Tobias Isenberg, and Ivan Viola. Computers & Graphics. DOI: https://doi.org/10.1016/j.cag.2025.104376 Link: https://www.sciencedirect.com/science/article/pii/S0097849325002171?via%3Dihub to analyze and understand molecular dynamics simulation trajectories of their designs. We use a multiscale multi-dimensional abstraction space, connect the representations to a projected conformational space plot of the structure’s temporal sequence, and thus enable experts to analyze the dynamics of their structural designs and, specifically, failure cases of the assembly. In addition, our time-dependent abstraction representation allows the biologists, for the first time in a smooth and structurally clear way, to identify and observe temporal transitions of a DNA-nano design from one configuration to another, and to highlight important periods of the simulation for further analysis. We realize SynopFrame as a dashboard of the different synchronized 3D spatial and 2D schematic visual representations, with a color overlay to show essential properties such as the status of hydrogen bonds. The linking of the spatial, schematic, and abstract views ensures that users can effectively analyze the high-frequency motion. We also categorize the status of the hydrogen bonds into a new format to allow us to color-encode it and overlay it on the representations. To demonstrate the utility of SynopFrame, we describe example usage scenarios and report user feedback. ### 2024 #### DiffFit: Visually-Guided Differentiable Fitting of Molecule Structures to a Cryo-EM Map Deng Luo, Zainab Alsuwaykit, Dawar Khan, Ondřej Strnad, Tobias Isenberg, Ivan Viola. IEEE Transactions on Visualization and Computer Graphics. DOI: https://doi.org/10.1109/TVCG.2024.3456404 Link: https://ieeexplore.ieee.org/document/10672541 We introduce DiffFit, a differentiable algorithm for fitting protein atomistic structures into an experimental reconstructed Cryo-Electron Microscopy (cryo-EM) volume map. In structural biology, this process is necessary to semi-automatically composite large mesoscale models of complex protein assemblies and complete cellular structures that are based on measured cryo-EM data. The current approaches require manual fitting in three dimensions to start, resulting in approximately aligned structures followed by an automated fine-tuning of the alignment. The DiffFit approach enables domain scientists to fit new structures automatically and visualize the results for inspection and interactive revision. The fitting begins with differentiable three-dimensional (3D) rigid transformations of the protein atom coordinates followed by sampling the density values at the atom coordinates from the target cryo-EM volume. To ensure a meaningful correlation between the sampled densities and the protein structure, we proposed a novel loss function based on a multi-resolution volume-array approach and the exploitation of the negative space. This loss function serves as a critical metric for assessing the fitting quality, ensuring the fitting accuracy and an improved visualization of the results. We assessed the placement quality of DiffFit with several large, realistic datasets and found it to be superior to that of previous methods. We further evaluated our method in two use cases: automating the integration of known composite structures into larger protein complexes and facilitating the fitting of predicted protein domains into volume densities to aid researchers in identifying unknown proteins. We implemented our algorithm as an open-source plugin (github.com/nanovis/DiffFit) in ChimeraX, a leading visualization software in the field. #### Nanomatrix: Scalable Construction of Crowded Biological Environments Ruwayda Alharbi, Ondřej Strnad, Tobias Klein, Ivan Viola. IEEE Transactions on Visualization and Computer Graphics. DOI: https://doi.org/10.1109/TVCG.2024.3411786 Link: https://ieeexplore.ieee.org/document/10552440 We present a novel method for the interactive construction and rendering of extremely large molecular scenes, capable of representing multiple biological cells in atomistic detail. Our method is tailored for scenes, which are procedurally constructed, based on a given set of building rules. Rendering of large scenes normally requires the entire scene available in-core, or alternatively, it requires out-of-core management to load data into the memory hierarchy as a part of the rendering loop. Instead of out-of-core memory management, we propose to procedurally generate the scene on-demand on the fly. The key idea is a positional- and view-dependent procedural scene-construction strategy, where only a fraction of the atomistic scene around the camera is available in the GPU memory at any given time. The atomistic detail is populated into a uniform-space partitioning using a grid that covers the entire scene. Most of the grid cells are not filled with geometry, only those are populated that are potentially seen by the camera. The atomistic detail is populated in a compute shader and its representation is connected with acceleration data structures for hardware ray-tracing of modern GPUs. Objects which are far away, where atomistic detail is not perceivable from a given viewpoint, are represented by a triangle mesh mapped with a seamless texture, generated from the rendering of geometry from atomistic detail. The algorithm consists of two pipelines, the construction-compute pipeline, and the rendering pipeline, which work together to render molecular scenes at an atomistic resolution far beyond the limit of the GPU memory containing trillions of atoms. We demonstrate our technique on multiple models of SARS-CoV-2 and the red blood cell. #### Volume conductor: interactive visibility management for crowded volumes Žiga Lesar, Ruwayda Alharbi, Ciril Bohak, Ondřej Strnad, Christoph Heinzl, Matija Marolt & Ivan Viola. The Visual Computer. DOI: https://doi.org/10.1007/s00371-023-02828-8 We present a novel smart visibility system for visualizing crowded volumetric data containing many object instances. The presented approach allows users to form groups of objects through membership predicates and to individually control the visibility of the instances in each group. Unlike previous smart visibility approaches, our approach controls the visibility on a per-instance basis and decides which instances are displayed or hidden based on the membership predicates and the current view. Thus, cluttered and dense volumes that are notoriously difficult to explore effectively are automatically sparsified so that the essential information is extracted and presented to the user. The proposed system is generic and can be easily integrated into existing volume rendering applications and applied to many different domains. We demonstrate the use of the volume conductor for visualizing fiber-reinforced polymers and intracellular organelle structures. #### Towards Differentiable Electron Microscopy Simulation: Methods and Applications Ngan Nguyen, Feng Liang, Dominik Engel, Ciril Bohak, Ondřej Strnad, Timo Ropinski, Ivan Viola. IEEE 17th Pacific Visualization Conference (PacificVis). DOI: https://doi.org/10.1109/PacificVis60374.2024.00016 We propose a new microscopy simulation system that can depict atomistic models in a micrograph visual style, similar to physical electron microscopy imaging results. This system is scalable, able to represent the simulation of electron microscopy of twenty complex viral particles, and synthesizes the image faster than previous approaches. Additionally, the simulator is differentiable in the deterministic and stochastic stages that form signal and noise representations in the micrograph. This allows for solving inverse problems by means of optimization and thus allows for the generation of microscopy simulations using parameter settings estimated from real data. We demonstrate this capability through two applications: (1) estimating the parameters of the modulation transfer function defining the detector properties of the simulated and real micrographs and (2) denoising the real data based on parameters optimized from the simulated examples using gradient descent. While current simulators do not support any parameter estimation due to their forward-only design, we show that the results obtained using estimated parameters are very similar to real micrographs. Additionally, we evaluate the denoising capabilities of our approach and show that the results are competitive with state-of-the-art methods. Denoised micrographs exhibit less noise in the tomographic reconstruction of tilt-series, facilitating visualization of microscopy tomography using direct volume rendering by reducing the visual dominance of noise. ### 2023 #### SpyDirect: a novel biofunctionalization method for high stability and longevity of electronic biosensors Keying Guo, Raik Grünberg, Yuxiang Ren, Tianrui Chang, Shofarul Wustoni, Ondrej Strnad, Anil Koklu, Escarlet Díaz‐Galicia, Jessica Parrado Agudelo, Victor Druet, Tania Cecilia Hidalgo Castillo, Maximilian Moser, David Ohayon, Adel Hama, Ashraf Dada, Iain McCulloch, Ivan Viola, Stefan T Arold, Sahika Inal. Advanced Science. DOI: https://doi.org/10.1002/advs.202306716 Electronic immunosensors are indispensable tools for diagnostics, particularly in scenarios demanding immediate results. Conventionally, these sensors rely on the chemical immobilization of antibodies onto electrodes. However, globular proteins tend to adsorb and unfold on these surfaces. Therefore, self-assembled monolayers (SAMs) of thiolated alkyl molecules are commonly used for indirect gold–antibody coupling. Here, a limitation associated with SAMs is revealed, wherein they curtail the longevity of protein sensors, particularly when integrated into the state-of-the-art transducer of organic bioelectronics—the organic electrochemical transistor. The SpyDirect method is introduced, generating an ultrahigh-density array of oriented nanobody receptors stably linked to the gold electrode without any SAMs. It is accomplished by directly coupling cysteine-terminated and orientation-optimized spyTag peptides, onto which nanobody-spyCatcher fusion proteins are autocatalytically attached, yielding a dense and uniform biorecognition layer. The structure-guided design optimizes the conformation and packing of flexibly tethered nanobodies. This biolayer enhances shelf-life and reduces background noise in various complex media. SpyDirect functionalization is faster and easier than SAM-based methods and does not necessitate organic solvents, rendering the sensors eco-friendly, accessible, and amenable to scalability. SpyDirect represents a broadly applicable biofunctionalization method for enhancing the cost-effectiveness, sustainability, and longevity of electronic biosensors, all without compromising sensitivity. ### 2022 #### Finding nano-Ötzi: cryo-electron tomography visualization guided by learned segmentation Ngan Nguyen, Ciril Bohak, Dominik Engel, Peter Mindek, Ondřej Strnad, Peter Wonka, Sai Li, Timo Ropinski, Ivan Viola. IEEE Transactions on Visualization and Computer Graphics. DOI: https://doi.org/10.1109/TVCG.2022.3186146 Cryo-electron tomography (cryo-ET) is a new 3D imaging technique with unprecedented potential for resolving submicron structural details. Existing volume visualization methods, however, are not able to reveal details of interest due to low signal-to-noise ratio. In order to design more powerful transfer functions, we propose leveraging soft segmentation as an explicit component of visualization for noisy volumes. Our technical realization is based on semi-supervised learning, where we combine the advantages of two segmentation algorithms. First, the weak segmentation algorithm provides good results for propagating sparse user-provided labels to other voxels in the same volume and is used to generate dense pseudo-labels. Second, the powerful deep-learning-based segmentation algorithm learns from these pseudo-labels to generalize the segmentation to other unseen volumes, a task that the weak segmentation algorithm fails at completely. The proposed volume visualization uses deep-learning-based segmentation as a component for segmentation-aware transfer function design. Appropriate ramp parameters can be suggested automatically through frequency distribution analysis. Furthermore, our visualization uses gradient-free ambient occlusion shading to further suppress the visual presence of noise, and to give structural detail the desired prominence. The cryo-ET data studied in our technical experiments are based on the highest-quality tilted series of intact SARS-CoV-2 virions. Our technique shows the high impact in target sciences for visual data analysis of very noisy volumes that cannot be visualized with existing techniques. #### GPU Accelerated 3D Tomographic Reconstruction and Visualization from Noisy Electron Microscopy Tilt-Series Julio Rey Ramirez, Peter Rautek, Ciril Bohak, Ondřej Strnad, Ivan Viola, Wolfgang Heidrich. IEEE Transactions on Visualization and Computer Graphics. DOI: https://doi.org/10.1109/TVCG.2022.3230445 We present a novel framework for 3D tomographic reconstruction and visualization of tomograms from noisy electron microscopy tilt-series. Our technique takes as an input aligned tilt-series from cryogenic electron microscopy and creates denoised 3D tomograms using a proximal jointly-optimized approach that iteratively performs reconstruction and denoising, relieving the users of the need to select appropriate denoising algorithms in the pre-reconstruction or post-reconstruction steps. The whole process is accelerated by exploiting parallelism on modern GPUs, and the results can be visualized immediately after the reconstruction using volume rendering tools incorporated in the framework. We show that our technique can be used with multiple combinations of reconstruction algorithms and regularizers, thanks to the flexibility provided by proximal algorithms. Additionally, the reconstruction framework is open-source and can be easily extended with additional reconstruction and denoising methods. Furthermore, our approach enables visualization of reconstruction error throughout the iterative process within the reconstructed tomogram and on projection planes of the input tilt-series. We evaluate our approach in comparison with state-of-the-art approaches and additionally show how our error visualization can be used for reconstruction evaluation. ### 2021 #### Molecumentary: Adaptable Narrated Documentaries Using Molecular Visualization David Kouřil, Ondřej Strnad, Peter Mindek, Sarkis Halladjian, Tobias Isenberg, M. Eduard Groller, Ivan Viola. IEEE Transactions on Visualization and Computer Graphics. DOI: https://doi.org/10.1109/TVCG.2021.3130670 We present a method for producing documentary-style content using real-time scientific visualization. We introduce molecumentaries, i.e., molecular documentaries featuring structural models from molecular biology, created through adaptable methods instead of the rigid traditional production pipeline. Our work is motivated by the rapid evolution of scientific visualization and it potential in science dissemination. Without some form of explanation or guidance, however, novices and lay-persons often find it difficult to gain insights from the visualization itself. We integrate such knowledge using the verbal channel and provide it along an engaging visual presentation. To realize the synthesis of a molecumentary, we provide technical solutions along two major production steps: (1) preparing a story structure and (2) turning the story into a concrete narrative. In the first step, we compile information about the model from heterogeneous sources into a story graph. We combine local knowledge with external sources to complete the story graph and enrich the final result. In the second step, we synthesize a narrative, i.e., story elements presented in sequence, using the story graph. We then traverse the story graph and generate a virtual tour, using automated camera and visualization transitions. We turn texts written by domain experts into verbal representations using text-to-speech functionality and provide them as a commentary. Using the described framework, we synthesize fly-throughs with descriptions: automatic ones that mimic a manually authored documentary or semi-automatic ones which guide the documentary narrative solely through curated textual input. #### Nanotilus: Generator of Immersive Guided-Tours in Crowded 3D Environments Ruwayda Alharbi, Ondřej Strnad, Laura R. Luidolt, Manuela Waldner, David Kouřil,Ciril Bohak, Tobias Klein, Eduard Groller, Ivan Viola. IEEE Transactions on Visualization and Computer Graphics. DOI: https://doi.org/10.1109/TVCG.2021.3133592 Immersive virtual reality environments are gaining popularity for studying and exploring crowded three-dimensional structures. When reaching very high structural densities, the natural depiction of the scene produces impenetrable clutter and requires visibility and occlusion management strategies for exploration and orientation. Strategies developed to address the crowdedness in desktop applications, however, inhibit the feeling of immersion. They result in nonimmersive, desktop-style outside-in viewing in virtual reality. This article proposes Nanotilus-a new visibility and guidance approach for very dense environments that generates an endoscopic inside-out experience instead of outside-in viewing, preserving the immersive aspect of virtual reality. The approach consists of two novel, tightly coupled mechanisms that control scene sparsification simultaneously with camera path planning. The sparsification strategy is localized around the camera and is realized as a multi-scale, multi-shell, variety-preserving technique. When Nanotilus dives into the structures to capture internal details residing on multiple scales, it guides the camera using depth-based path planning. In addition to sparsification and path planning, we complete the tour generation with an animation controller, textual annotation, and text-to-visualization conversion. We demonstrate the generated guided tours on mesoscopic biological models - SARS-CoV-2 and HIV. We evaluate the Nanotilus experience with a baseline outside-in sparsification and navigational technique in a formal user study with 29 participants. While users can maintain a better overview using the outside-in sparsification, the study confirms our hypothesis that Nanotilus leads to stronger engagement and immersion. ### 2020 #### Modeling in the time of COVID-19: Statistical and rule-based mesoscale models Ngan Nguyen, Ondřej Strnad, Tobias Klein, Deng Luo, Ruwayda Alharbi, Peter Wonka, Martina Maritan, Peter Mindek, Ludovic Autin, David S Goodsell, Ivan Viola. IEEE Transactions on Visualization and Computer Graphics. DOI: https://doi.org/10.1109/TVCG.2020.3030415 We present a new technique for the rapid modeling and construction of scientifically accurate mesoscale biological models. The resulting 3D models are based on a few 2D microscopy scans and the latest knowledge available about the biological entity, represented as a set of geometric relationships. Our new visual-programming technique is based on statistical and rule-based modeling approaches that are rapid to author, fast to construct, and easy to revise. From a few 2D microscopy scans, we determine the statistical properties of various structural aspects, such as the outer membrane shape, the spatial properties, and the distribution characteristics of the macromolecular elements on the membrane. This information is utilized in the construction of the 3D model. Once all the imaging evidence is incorporated into the model, additional information can be incorporated by interactively defining the rules that spatially characterize the rest of the biological entity, such as mutual interactions among macromolecules, and their distances and orientations relative to other structures. These rules are defined through an intuitive 3D interactive visualization as a visual-programming feedback loop. We demonstrate the applicability of our approach on a use case of the modeling procedure of the SARS-CoV-2 virion ultrastructure. This atomistic model, which we present here, can steer biological research to new promising directions in our efforts to fight the spread of the virus. ### 2018 #### CAVER Analyst 2.0: Analysis and Visualization of Channels and Tunnels in Protein Structures and Molecular Dynamics Trajectories Adam Jurčík, David Bednář, Jan Byška, Sérgio Manuel Marques, Katarína Furmanová, Lukáš Daniel, Piia Pauliina Kokkonen, Jan Brezovský, Ondřej Strnad, Jan Štourač, Antonín Pavelka, Martin Manak, Jiří Damborský a Barbora Kozlíková. BMC Bioinformatics. DOI: https://doi.org/10.1093/bioinformatics/bty386 Studying the transport paths of ligands, solvents, or ions in transmembrane proteins and proteins with buried binding sites is fundamental to the understanding of their biological function. A detailed analysis of the structural features influencing the transport paths is also important for engineering proteins for biomedical and biotechnological applications. CAVER Analyst 2.0 is a software tool for quantitative analysis and real-time visualization of tunnels and channels in static and dynamic structures. This version provides the users with many new functions, including advanced techniques for intuitive visual inspection of the spatiotemporal behavior of tunnels and channels. Novel integrated algorithms allow an efficient analysis and data reduction in large protein structures and molecular dynamic simulations. ### 2014 #### CAVER Analyst 1.0: Graphic tool for interactive visualization and analysis of tunnels and channels in protein structures Barbora Kozlíková, Eva Šebestová, Vilém Šustr, Jan Brezovský, Ondřej Strnad, Lukáš Daniel, David Bednář, Antonín Pavelka, Martin Manak, Martin Bezděka, Petr Beneš, Matúš Kotry, Artur Wiktor Gora, Jiří Damborský a Jiří Sochor. BMC Bioinformatics. DOI: https://doi.org/10.1093/bioinformatics/btu364 The transport of ligands, ions or solvent molecules into proteins with buried binding sites or through the membrane is enabled by protein tunnels and channels. CAVER Analyst is a software tool for calculation, analysis and real-time visualization of access tunnels and channels in static and dynamic protein structures. It provides an intuitive graphic user interface for setting up the calculation and interactive exploration of identified tunnels/channels and their characteristics. #### Detection of Intramolecular Tunnels Connecting Sequence of Sites in Protein Structures Ondřej Strnad, Barbora Kozlíková a Jiří Sochor. Advances in Intelligent Systems and Computing. DOI: https://doi.org/10.1007/978-3-319-07581-5_9 Proteins are essential for functioning of all living organisms and studying their inner structure and functions has been of a high importance. Many studies concentrated on detection of various inner structures inside macromolecules (e.g. tunnels, channels, pores) which play an essential role in the functioning of a large number of proteins. Here we present a novel approach to a detection of intramolecular tunnels. These pathways may facilitate the transport of reaction intermediates among buried active sites. The results obtained by the proposed algorithm were compared to intramolecular tunnels whose presence in given structures is already known. The algorithm is able to also identify other inner structures, such as channels or pores. #### Geometrical Detection of Pathways in Protein Structures Leading Among More Binding Sites Ondřej Strnad, Vilém Šustr, Barbora Kozlíková and Jiří Sochor. Proceedings of BIOTECHNO 2014. In this paper, we present a novel algorithm for the detection of pathways connecting two or more specific user defined binding sites, which are deeply buried in a protein macromolecule. These pathways can play an important role in the protein reactivity and overall behavior. However, our new algorithm can be generalized and used for computation of pathways inside an arbitrary set of spheres in three-dimensional space, leading through an ordered set of user-defined sites. Our approach is based on the localized Voronoi diagram approach and the Delaunay triangulation. The greatest benefit of our approach is its independence on the size of the input data set. This is achieved by using only a subset of all atoms in the macromolecule in each phase. This substantially reduces the size of the processed space. The method can also be utilized for determination whether pathways wide and straight enough exist among determined binding sites. This information then serves as the guideline for assessing the migration of products of chemical reaction between these binding sites. #### Visualizing Movements of Protein Tunnels in Molecular Dynamics Simulations. Barbora Kozlíková, Adam Jurcík, Jan Byska, Ondrej Strnad, Jirí Sochor. Eurographics Workshop on Visual Computing for Biology and Medicine. DOI: https://doi.org/10.2312/vcbm.20141188 Analysis and visualization of molecules and their structural features help biochemists and biologists to better understand protein behavior. Studying these structures in molecular dynamics simulations enhances this understanding. In this paper we introduce three approaches for animating specific inner pathways composed of an empty space between atoms, called tunnels. These tunnels facilitate the transport of small molecules, water solvent and ions in many proteins. They help researchers understand the structure-function relationships of proteins and the knowledge of tunnel properties improves the design of new inhibitors. Our methods are derived from selected tunnel representations when each stresses some of the important tunnel properties — width, shape, mapping of physico-chemical properties, etc. Our methods provide smooth animation of the movement of tunnels as they change their length and shape throughout the simulation. ### 2013 #### Real-time visualization and exploration of protein empty space with varying parameters Ondřej Strnad, Barbora Kozliková, Vilém Šustr, Jiří Sochor. International Journal on Advances in Life Sciences. Long-term research in the area of protein analysis proved the importance of an empty space situated inside these macromolecular structures. This empty space influences the protein function, characteristics or reactivity. Many algorithms enabling computation of these empty spaces (or voids) have been published and their results were evaluated by protein engineers to confirm their chemical relevance. However, not all detected voids inside protein are of the same importance. Thus, the examination and assessment of all voids must follow to reveal the important ones. In this phase the visual representation of voids is very valuable and substantially decreases the time spent in this evaluation phase. In this paper we present an extension of the algorithm for the visualization and further evaluation of protein voids in real-time. The userdriven approach enables to compute and display empty space that satisfies the input parameters instantly. The values of these parameters can be changed by the user anytime and the changes are immediately displayed and prepared for further exploration. Our improvements involve an exclusion of selected atom or group of atoms (ligands, ions) from the computation, which can change the size and shape of the detected void. Another improvement is related to the detection of the binding site which is usually located in one of the largest voids. So the algorithm suggests and visually separates (by different coloring) the largest void of given area. Several improvements were also made in the field of real-time exploration – currently the interaction on large structures is fluent. In consequence, the current version of the algorithm provides the biochemists with very adjustable and precise algorithm for detection of inner voids in a user-defined region of protein structures. ### 2012 #### CAVER 3.0: A Tool for the Analysis of Transport Pathways in Dynamic Protein Structures Eva Chovancová, Antonín Pavelka, Petr Beneš, Ondřej Strnad, Jan Brezovský, Barbora Kozlíková, Artur Wiktor GORA, Vilém Šustr, Martin Klvaňa, Petr Medek, Lada Biedermannová, Jiří Sochor a Jiří Damborský. PLoS Computational Biology. DOI: https://doi.org/10.1371/journal.pcbi.1002708 Tunnels and channels facilitate the transport of small molecules, ions and water solvent in a large variety of proteins. Characteristics of individual transport pathways, including their geometry, physico-chemical properties and dynamics are instrumental for understanding of structure-function relationships of these proteins, for the design of new inhibitors and construction of improved biocatalysts. CAVER is a software tool widely used for the identification and characterization of transport pathways in static macromolecular structures. Herein we present a new version of CAVER enabling automatic analysis of tunnels and channels in large ensembles of protein conformations. CAVER 3.0 implements new algorithms for the calculation and clustering of pathways. A trajectory from a molecular dynamics simulation serves as the typical input, while detailed characteristics and summary statistics of the time evolution of individual pathways are provided in the outputs. To illustrate the capabilities of CAVER 3.0, the tool was applied for the analysis of molecular dynamics simulation of the microbial enzyme haloalkane dehalogenase DhaA. CAVER 3.0 safely identified and reliably estimated the importance of all previously published DhaA tunnels, including the tunnels closed in DhaA crystal structures. Obtained results clearly demonstrate that analysis of molecular dynamics simulation is essential for the estimation of pathway characteristics and elucidation of the structural basis of the tunnel gating. CAVER 3.0 paves the way for the study of important biochemical phenomena in the area of molecular transport, molecular recognition and enzymatic catalysis. The software is freely available as a multiplatform command-line application at http://www.caver.cz. ### 2011 #### Computation of dynamic channels in proteins Petr Beneš, Petr Medek, Ondrej Strnad, Jiri Sochor. Bioinformatics. In this paper, we propose a new method which considers the movement of a protein molecule as a whole for the computation of so called dynamic channels in a molecular dynamics trajectory. The method is based on maximizing the information about the empty space over time and is built on basic computational geometry principles. The dynamic channels highlight pulsing and flexible parts of the molecule. It is believed that such parts allow a ligand to pass into or out from the active site. The method was tested on real protein data and the results indicate that it presents new information about the molecule. #### New path planning method for computation of constrained dynamic channels in proteins Petr Beneš, Ondřej Strnad, Jiří Sochor. WSCG 2011: Full Papers Proceedings: The 19th International Conference in Central Europe on Computer Graphics, Visualization and Computer Vision in cooperation with EUROGRAPHICS Association. Collision-free paths in the geometric model of a protein molecule reveal various dependencies between the structure of the molecule and its function. The paths which connect a biochemically important part of the protein molecule with the surface of the molecule can serve as egression or access paths for small molecules which react in the active site. The geometric method introduced in this paper is designed to compute such paths in the dynamic models of protein molecules. The paths have to satisfy additional constraints such as valid flow of time which allows us to distinguish between access and egression paths, minimum width and others. Possibly, the method may be used not only for protein molecules but also for similar environments with high density of spherical obstacles. The method was tested on real protein data and the results indicate that if there is a path present, it is detected by our method.