X

The challenge in this case study is the need for efficient and effective processing, management, and analysis of omic and phenotypic data in translational research activities. These activities are crucial for characterizing and profiling patients using omic technologies to understand their response to new therapies, stratify patients for trials, or search for new disease biomarkers. The existing in-house software solutions such as tranSMART, while useful, require enhancement and expansion to fully optimize the process of translational research. The data quality and curation is critical to making the right scientific conclusions, and the process of loading omics and clinical sample annotations (metadata) into tranSMART can be time consuming and expensive.

Novozymes, a world leader in biological solutions, uses bacterial and fungal hosts to manufacture bulk enzymes for a variety of industries. Host cell proteins (HCPs) are a major source of contamination that can adversely affect product stability and performance. The company needed to automate and streamline each stage of their LC-MS approach to keep pace with the large number of samples that their lab is required to analyze. Implementing robotic sample preparation, a shortened HPLC separation step, and high-speed MS acquisition increased sample throughput but generated large amounts of complex data requiring time-consuming analysis and review. The company also needed a fast and efficient way to ingest results into their corporate data lake and provide fast and efficient ways to share relevant information to key stakeholders and decision makers.

Biopharmaceutical companies are facing the challenge of managing and utilizing their R&D data, which is often siloed within different functions of the company. This siloed data structure hinders the effective federation of distributed data, which is crucial for increasing clinical trial success rates. The industry is recognizing the value of sharing and reusing data for multiple analyses, which requires breaking down these data silos. Furthermore, the data generated in biopharmaceutical R&D is complex and originates from various sources, making it crucial to implement the FAIR (Findable, Accessible, Interoperable, and Reusable) principles for scientific data management and stewardship. However, making data FAIR is not an easy task and requires the right infrastructure that can handle the data volume and privacy of patient-level information.

AstraZeneca, a global biopharmaceutical company, was facing challenges in automating data processing and increasing the quality of results. The company was using Acoustic Droplet Ejection (ADE) technology for high-throughput screening (HTS), which introduced samples into a mass spectrometer. However, the increase in throughput brought by Acoustic Mist Ionization Mass Spectrometry (AMI-MS) technology resulted in large amounts of data that needed efficient processing. The main challenges included eliminating laborious and potentially error-inducing manual interventions, guaranteeing data quality and traceability, adapting data processing to different targets and screening strategies, obtaining insightful visualization to speed assay development, and integrating MS data into screening data analysis platforms.

Immatics is a clinical-stage biopharmaceutical company that focuses on the development of novel therapies to treat cancer patients with solid tumors. The company employs two innovative proprietary technologies, XPRESIDENT® and XCEPTOR®, to develop T cell receptor-based immunotherapies. The TCR-related R&D workflow at Immatics is very sophisticated, involving hundreds of individual steps and iterations to generate new molecules and cell lines, engineer and optimize the therapeutic modalities, and constantly test and monitor them to see if they are fit-for-purpose. Each step generates critically important data that needs to be captured, processed, analyzed, and made available to various R&D teams. The complex data stream includes all molecule, sequence, and sample details, assay, and analytics values, as well as decisions such as the CDR combinations to be used to engineer distinctive TCR bispecific molecules. This data must be integrated, stored, related, and interpreted in real time to support the day-to-day activities of the various Immatics’ R&D teams.

MorphoSys, a late-stage biopharmaceutical company, was facing challenges due to its rapid growth. The increase in molecule and sample throughput led to a significant amount of R&D data, which was difficult to capture, process, and interpret. The scale and throughput of their operations necessitated the use of an enterprise workflow platform to manage and streamline the growing number of their discovery programs. The goal of the new system was to make MorphoSys’ sophisticated R&D processes more efficient and to facilitate handovers between various R&D teams and functions, including screening, molecular biology, engineering, expression, purification, and analytics. The new system also needed to be able to handle the increasing amount of data and samples produced by external partners, such as CROs, which had to be integrated into MorphoSys’ R&D process.

Bayer Pharmaceuticals obtained an exclusive license from Ionis Pharmaceuticals to develop new oligonucleotide-based treatments for a wide range of medical conditions. To support release testing and the imminent clinical trials of an advanced drug candidate, the Bayer Analytical Development Team for Biologics was tasked with implementing a validated drug substance and drug product release ID test that required establishing new analytical methods, standard operating procedures, and organizational responsibilities. The tight go-live timeline and anticipated costs of any delay led Bayer to work together with our long-term partner Genedata to develop a confirmatory identification (ID) workflow of oligonucleotide species by mass spectrometry so that an approved quality control (QC) could be implemented and validated by Bayer’s Quality Assurance (QA) team. The main challenges were implementing an oligonucleotide analytics workflow, satisfying GMP requirements, and meeting a wide range of new analytical demands.

Bayer Pharmaceuticals was facing several challenges in their lab. They were developing new methods to characterize novel classes of biopharmaceuticals, which was compounded by the novel nature of the analytes, for which no established MS-based protocols were available. Interpreting results and iteratively optimizing experimental and analytical methods for such molecules required a significant amount of area expertise in MS-based characterization of biotherapeutics. They also needed to deliver productive output in a timely manner to meet the needs of their internal customers involved in time-critical projects and maximize Bayer’s return on investment in new infrastructure. The novel nature of their analytes and protocols and the limitations of the available software made interpretation of results far from straightforward. When troubleshooting a complex process issue, it was extremely helpful to have not only domain expertise, but also to have a comprehensive and unbiased overview of all data at all stages of processing. Any solution that they implemented had to be scalable and able to handle high sample throughputs. Using the available software, data processing and the production of targeted reports were laborious and time-consuming processes that required frequent manual interventions.

Pfizer, a global biopharma giant, faced the challenge of aligning workflows and centralizing data for more than 200 scientists located at 6 sites around the world. The company's diverse and growing biopharma R&D operations required a central system to share data and align large-molecule R&D processes. With R&D teams spread across the globe, vast amounts of data were being collected in silos, making it difficult for teams to share information. This impeded collaboration and led to process inefficiencies. After an aborted attempt with a develop-as-you-go approach, Pfizer decided to evaluate the market in search of a commercial enterprise data management system for their large-molecule R&D groups. The desired system would need to act as a backbone platform and central repository for Pfizer’s biological R&D workflows and include tools able to capture specific instrumental data, workflows, analyses, and each group’s unique contribution to the overall biological drug discovery process at Pfizer.

The Microbial Process Development Group at Merck KGaA was tasked with developing and producing recombinant proteins expressed in Escherichia coli and Pichia pastoris. They used mass spectrometry (MS) as a routine tool for supporting process development in recombinant protein production. However, routine use of Multi-attribute method (MAM) in this environment meant overcoming various scientific, technological, and methodological challenges. These challenges included managing large amounts of data, producing unbiased audited results, and meeting process validation requirements. The overall production process for recombinant proteins involves multiple processing steps that are driven by defined process parameters. The characteristics of the resulting protein product must be such that the final drug substance is both safe and efficient. Studying the desired protein characteristics, then defining, monitoring, and managing critical quality attributes (CQAs) is key to success.

Zeleros, a startup developing an innovative hyperloop system to revolutionize ground travel, was facing challenges with its local computing capabilities. The company's local computing capabilities lacked the flexibility to accommodate the changing requirements for simulations and modeling. The on-premises computing capacity was fully occupied by a single simulation, causing significant impediments to productivity. The team was also losing a significant amount of time benchmarking to find the appropriate HPC instance for each simulation.

Boundary Layer Technologies is a pioneering company that designs small hydrofoil cargo ships aimed at disrupting the transportation industry. The company's goal is to create a new class of vehicle that could revolutionize global trade. However, creating new technology is complex. A major challenge they faced was creating simulations that would account for free surface – the boundary between air and water – and cavitation effects on wing optimization. They also needed services that could handle the scale of their designs, as they were creating meshes that took hundreds of gigs of RAM that they could not fit on a local desktop.

ARC, an innovative aerospace company that 3D prints metal rocket engines, was facing a bottleneck in their product development pipeline due to the lack of scalability in computing resources and agility in resource diversity. Their local on-premise HPC system with 50-128 cores at 80% utilization was not sufficient to run the large number of simulations required for their product development. This severely limited ARC’s simulation-throughput and design of experiments (DOE), delaying their potential time-to-market. Faced with the urgent demand for more computing resources, ARC had to decide between investing in a static on-premise HPC system or moving to a cloud-enabled HPC system.

Pinnacle Engines, a company developing and commercializing an ultra-efficient engine architecture, needed to run a wide variety of simulations, including the analysis of a four-stroke internal combustion engine using the computational fluid dynamics (CFD) tool, CONVERGE by Convergent Science. The design of an engine with 30-50% fuel efficiency required many simulations, constant iterations, and the compute power to accomplish all this within project timelines. Running tens to hundreds of models simultaneously required extensive compute resources and access to a large pool of simulation software licenses, equating to large up-front costs and personnel investments. Even with a powerful internal processing cluster, the ability to run parallel jobs was significantly limited.

Exponent, a company providing a wide variety of engineering and scientific consulting services, was facing challenges due to the increasing technological complexity of their clients' demands. Their clients, who are building next-generation consumer electronics, energy, healthcare, and automotive products, were increasingly relying on Exponent’s expertise in computational fluid dynamics (CFD) and thermal management. However, the needs of these projects quickly outpaced the computational capacity of Exponent’s internal HPC hardware. The team needed a solution that could be right-sized for each client and simple enough for new engineers joining the team to learn and use quickly.

AGC, a global leader in glass, ceramic, chemicals, and electronics materials manufacturing, faced several challenges related to workflow and how users actively work with simulation tools and data. Sophisticated microstructure geometry designs needed to be modeled from the molecular-level up to the system level. These simulation models are very computationally-intensive and need to be performed continuously and across each step in the design and manufacturing process. Managing complex processes and workflows was further complicated by software updates. AGC uses a diverse suite of applications like LAMMPS, STAR-CCM+®, and COMSOL Multiphysics® to model different types of physics, and it was difficult to keep up with all of the software installation updates and maintenance. As a global enterprise with the highest security standards and requirements, AGC needed the solution to comply with the strictest IT policies, yet also simplify how users work with data and software tools and obtain resources. Other challenges included integrating security from the on-premise facilities to the cloud, minimizing data movement across the wide-area network (WAN) for hybrid cloud workflows, software licensing, accommodating complex, multi-step workflows, administration of policies, methodologies, software, and management tools, and future compatibility to enable breakthrough in areas such as machine learning and deep learning.

Richard Childress Racing (RCR) is a successful NASCAR team that designs and builds its race cars from the ground up. Over the past 12 years, RCR has invested significantly in computational fluid dynamics (CFD) to develop a deeper understanding of the aerodynamics of their cars, evaluate new aerodynamic concepts, and analyze phenomena not modeled in the wind tunnel. However, CFD is a compute-intensive process, and RCR's on-premises resources were limited. They needed to augment their capacity to build larger models with a resolution high enough to precisely understand the intricate flow details that affect the car’s speed on the track.

The global automotive manufacturer was already using multiple HPC cloud services providers to power its digital engineering. However, the company wanted to further streamline its operations by moving its engineering workstations to the cloud. The challenge was to do this without incurring unnecessary costs for idle workstations. Traditional physical workstations were limited in their compute capacity and required significant time to boot up and configure. The company wanted to avoid the costs of cloud compute services when they weren't using the workstations, which was a significant portion of the total time to complete an experiment.

In the automotive business, major automakers send out requests for bids to equipment manufacturers when they need a new part. These requests typically have very short timelines, requiring a response in just a few weeks. The potential business from these contracts—often for millions of parts—is central to the success of auto parts manufacturers. It is essential that they can quickly engineer new equipment designs that perform to specifications, are reliable, and can be sold at a profitable margin. One of Rescale’s customers faced this exact challenge. The R&D teams for this auto equipment manufacturer are at the center of its efforts to develop new designs to win contracts with the major automakers. So it is imperative the company does anything it can to better power its engineering and design process. Without the necessary compute power, the automotive manufacturer was limited in its abilities to fully research and test its new designs before making a bid.

TEN TECH LLC, an ITAR certified company, often services the highly confidential and secure sector of military and government organizations. They require a flexible yet highly secure platform for running models for their military and defense contractors. National and international clients and deadlines mean that they need access to extensive compute resources with the capability of handling large CFD and FEA simulations. A public model TEN TECH LLC recently ran was an analysis studying the effects of hurricane-force winds on large telescope arrays, including stress, deflection, and aeroelasticity. With their local server used for other projects, TEN TECH LLC needed an immediate solution capable of handling their compute-intensive project quickly.

Liberty University's engineering program, one of the fastest growing schools within the university, required computation-intensive R&D and faced challenges providing adequate high performance computing (HPC) resources. The HPC software and hardware required varied greatly across a wide range of engineering programs, including civil, electrical, industrial, mechanical, and computer engineering. Combining that variation with the ongoing need for support posed a challenge for the IT team, so they began to search for a solution that could quickly scale to the needs of the school. They had concerns around cost and their ability to deliver it quickly. When they explored their specific requirements they knew they were short on personnel to implement and maintain a solution on-premises, so they explored options for cloud HPC.

Boom Supersonic, an aerospace startup, is designing a supersonic passenger jet that will revolutionize business travel. However, creating a technological breakthrough in the aerospace sector traditionally requires billions of dollars in R&D, a large engineering staff, extensive wind tunnel testing, and many years of development. Advanced fluid-flow and mechanical stress simulation tools have reduced physical testing requirements and costs, but they require costly dedicated high performance computing (HPC) resources to be effective. As a small startup with limited funding but outsized dreams, Boom turned to the public cloud because the upfront costs of building their own on-premise HPC cluster were cost-prohibitive.

Nissan, a global full-line vehicle manufacturer, was facing challenges with its on-premise High Performance Computing (HPC) systems. The company was limited by fundamental aspects of on-premise computing, such as limited electric power, high total cost, and data center utilization challenges. Their on-premise HPC systems were constrained by the initial hardware and software specifications, were complex to operate, and struggled to handle high-demand (peak) loads. These inherent on-premise problems threatened Nissan's innovation, market leadership, agility, and time-to-market.

DENSO, a leading automotive and Fortune 500 company, operates globally in 35 countries and regions around the world and manufactures a wide variety of components. In its non-automotive business, DENSO is working to industrialize smart agriculture using factory automation and sensing technologies. Within the thermal management systems team, a core businesses at DENSO, new initiatives are underway to respond to electric vehicles and automated driving. Product development for thermal systems in electric vehicles involves the development of a key element of the thermal system, the product development in the thermal systems requires unconventional technologies and methods. It is essential to consider a variety of factors when undertaking new designs. Because of the limitations of manual checking through prototyping, simulation is becoming even more important to improve development efficiency. Mr. Ogawa has championed the development and promotion of simulation methods and tools in the Heat Exchanger R&D Division, which develops heat exchangers used in air conditioners and radiators, key components of thermal systems, and he describes the difficulties of utilizing simulation.

RWDI, a multidisciplinary engineering consulting firm, was facing a challenge due to its business growth. They had outgrown their on-premises High Performance Computing (HPC) environment and needed a solution that could handle their increasing workload. They had recently won a large wind speed mapping project across the Middle East, which required analyzing 30 years of atmospheric data across the entire region in two months. This required a million core-hours, significantly more than their on-premises system capacity. Due to the short project timeline, expanding their on-premises HPC was not feasible. Building out their own software and middleware infrastructure to be able to burst to the public cloud on their own was also not an option. They needed a turnkey hybrid cloud solution to expand their existing capacity.

Engineering world-renowned transportation requires high powered computing resources and access to industry-leading software. When internal capacity is reached, and there are tight project deadlines, Trek turns to Rescale to instantly expand their resources. In a recent analysis, Trek engineers ran a complex simulation using CFD tool Star-CCM+ by CD-adapco and optimizer HEEDS by Red Cedar Technology in a coupled analysis on Rescale’s cloud simulation platform. The goal was to study varying bicycle drafting methods to introduce new angles to the existing analysis method.

Specialized Bicycle Components, a leading manufacturer of performance bicycles, was facing constraints in their internal computing resources which were slowing down their product design cycles. The company's R&D teams were experiencing slow solve speeds which were limiting their ability to innovate and create new product designs. The company operates a proprietary wind tunnel for performance testing, but they also employ computer-aided engineering. However, their existing simulation workstations were becoming insufficient for the growing high performance computing needs of their Road Bike R&D team.

Cambridge Consultants, a world leader in disruptive innovation and technology-based consulting, was facing challenges in streamlining their lab operations due to the high turnover of projects. Their labs needed to be flexible and capable of developing and executing robust new protocols quickly. They required a scalable, robust tool that could provide end-to-end connectivity of devices and data to run protocols effectively, particularly complex multi-factorial experiments with many varying conditions. One area where they needed significant improvement was in the rapid and flexible development of liquid handling automation workflows. They needed a solution that could handle high throughput, fast execution speed, custom sample handling steps, robust sample traceability, excellent reproducibility, and minimized execution costs.

CSL set a goal of doubling the throughput of experimentation via robotics, automation & digital connectivity without a significant increase in headcount by 2030. The CSL Purification Team was looking for a solution that would enable scientists to plan and execute micro-scale purification experiments without the need to write code, ensuring their scientists could increase their throughput while improving the quality of their results. The team faced obstacles such as low uptake of automation and complicated scripting, low device utilization in lab environments, and manual management of experimental workflows & associated data. The team required a tool that would help them standardize their workflow design, execution, and data analysis processes for micro-scale purification experiments and easily transfer knowledge from one site to the next.

InterContinental Hotels Group (IHG) utilizes four different public clouds due to the size and complexity of their IT infrastructure. They wanted the flexibility of multiple public clouds to use the optimal environment for each workload and to protect themselves against price increases and instabilities in any one public cloud. They had been using VMware's vRealize Automation for managing VMware servers, but it did not provide support for the public clouds that they needed. Furthermore, they found vRA to be extremely time-consuming to maintain, with upgrades requiring a multi-month process and a large professional services cost. The process of installing multi-server, multi-tier apps on any one of these public clouds was onerous and required the administrator to use multiple different interfaces.