More than ever before enterprises have access to endless amounts of Data within various silos and systems, and to streamline this data collection is only one aspect to success, but making sense of it all and simplifying the use case creation is where maximum value is achieved. In this session with Gilead, in partnership with Thoughtworks, we will walk through the benefits of establishing Data as a Product & Data Mesh Principles as the foundation of their design approach, which enabled them to build a scalable enterprise-wide data platform “Gilead DnA”,  where data engineers and researchers can benefit from a secure self-service environment for their data processing needs. We will walk through the logical progression, from current state legacy processes, data quality challenges, and a lack of confidence in a cohesive strategy for data sharing and governance across the enterprise that continued to impede value realization, to their future state success of objectives and value we have achieved. 

This presentation will explain how Lakehouse in a Box is a service that simplifies the user experience for data producers and consumers by abstracting engineering complexities and eliminating redundant manual activities. It provides ready-to-use producer/consumer platforms and includes best practices and accelerators from both AWS and BMS, leveraging services such as Infrastructure as Code (IaC), Service Catalog, and Self-Service Infrastructure Portal. IaC is a key component of LHIAB, allowing for the management and provisioning of data lake infrastructure using code. This approach ensures consistency, repeatability, and scalability of the infrastructure, and involves defining the infrastructure requirements in code that can be version-controlled, tested, and deployed using automation tools. CloudFormation is typically used to define the infrastructure requirements in code, enabling the infrastructure to be defined as a set of declarative statements that can be version-controlled and tested like any other code. With IaC, data lake infrastructure requirements can be defined in a single codebase, which can be used to provision and manage the infrastructure across multiple environments. This helps to ensure consistency and reduces the risk of configuration drift. Automation tools can be used to automate the provisioning and management of the data lake infrastructure, reducing manual errors, and improving efficiency. For example, new resources can be provisioned automatically, existing resources can be updated, and routine maintenance tasks can be performed. 

The multi-tenant NGS LIMS developed at Genentech is a web application that can support multiple departments to manage their NGS workflow and data. The software integrates with existing company-wide sample registries and data analysis pipelines, acts as a crucial link between sample metadata and NGS data analysis results, and contributes to FAIR NGS data. The multi-tenant architecture also achieves cost-savings in software development, deployment, and maintenance.

Digital labs leveraging ML and AI algorithms will become commonplace in the future. Labs of the future will need to automate data management tasks such as data movement and sharing, to access diverse compute resources, and to fluidly cross authorization boundaries. Using AI-directed control of self-driving laboratories as an example, we will highlight how Globus can accelerate your move to modernize your infrastructure by enabling efficient and extensible automated task orchestration.

Our Omics start2end initiative enhances omics data accessibility and utility in biomedical research. By adhering to FAIR principles, it broadens data applications across studies. The platform streamlines tasks like registration and analysis, benefiting data scientists with consistent access. Upstream metadata quality has notably improved. Community adoption of these workflows and visualizations steadily grows, indicating a shift towards standardized, impactful methodologies.

This talk explores the importance of creating an integrated "Lab Automation as a Service" platform that can be used to accelerate the digitalization process of Pharma R&D. We will go over the design process of such a platform, how to create and cultivate digital innovation across the your R&D organization and dive into a series of success stories of how we are making a positive impact on Roche's R&D productivity.

When binding data for a target is not available, limited, or biased, most AI drug discovery companies are not well-equipped to deliver novel, viable starting points for optimization and/or leads. This is especially true for therapeutic targets in inflammation, where there is a narrow window of acceptability for ADME-tox properties due to the need for chronic dosing. To tackle these challenges, biases towards knowns need to be removed from every part of the design-build-test-learn cycle. DeepCure is removing limitations to novelty and diversity in each step, with AI tools that give a framework for identifying novel binding opportunities and design tools that allow drug design in novel chemical space while taking into account desired ADME-tox profile and synthesizability requirements. Accessing novel chemical space is made possible by a robotic automated custom synthesis system that allows for the reproducible multi-step synthesis of molecules from 100+ different reaction types.

Generate Biomedicines is a clinical stage therapeutic drug company that is pioneering an ML-powered generative biology platform that integrates digital and hardware automation to increase efficiency and decrease cycle times. Here, I will discuss how we tackle challenges posed by the flexibility, diversity, and fast pace of our R&D platform to feed data and metadata into Generate’s FAIR database.

CMC process development is facing multiple challenges from the increasing number of projects to higher agility, cost reductions, data integrity & accelerated development timelines. Data generated during drug development is only partially leveraged to further improve our processes or use it for predictive modeling. Building on a multi-year CMC digital transformation program, we embarked on digitizing end-to-end workflows across multiple platforms & lab families. We focused on three key areas: Operating Model, Value Delivery, and Scaling. We will share progress and challenges in each of these areas.

Advances in analytical technology combined with robotic high-throughput cell culture systems have resulted in tremendous volumes of data. To address these growing scientific and digital needs, the end-to-end cell culture digital transformation team has built automated data pipelines that enable near real-time data flow for scientists. Successful deployment has already occurred in Framingham to the automated HiTMaP (High-Throughput Mammalian Platform) lab with future rollouts to the hundreds of cell culture scientists at Sanofi.

This review covers advanced techniques for neural network characterization and explores emerging technologies for future insights. The presentation covers notable advances in brain organoid research, highlighting core measurements that companies like DBC focus on for drug development. Given the brain's electrical nature, understanding patterns of electrical activity in animal models and in organoids offers insights into real brain function. Managing the complex data from these measurements is also a significant challenge, and the presentation will convey the importance of novel data analysis methods.

The success of any laboratory hinges on the seamless integration and efficient operation of its instruments and automation workflows. We will present an innovative, real-time monitoring system for lab instruments that is vendor agnostic of specific devices.  Accessible from any web browser and displayed on TV dashboards throughout the lab, this system facilitates improvements in lab productivity and throughput. It provides a 'single pane of glass' view of lab instrument operations, enhancing visibility and control. The solution is based on a single enterprise platform, which provides easy scalability, IT manageability, and facilitates quick implementation—we call this “Lab Health TV."

In an era marked by technological advancements, this session explores the evolution of healthcare from centralized laboratories to decentralized home-based diagnostics. Join AI strategy consultant Talia Grace and Vital Biosciences CEO and co-founder Vasu Nadella as they seek to unravel the implications of this paradigm shift, from the empowerment of patients to the reimagining of care delivery.

Decisions affecting lab PCs are often touted being made using a “risk-based” approach. This involves gathering point-in-time data as well as various stakeholders to discuss and agree on the analyzed results. We will present our proof-of-concept that captures risk events, and calculates risk-scoring using real-time monitoring and defined weights based on the Six Sigma Failure Mode Effects Analysis methodology (FMEA) framework. This will allow teams to identify when technology is “increasing in risk” against baselines as well as where the overall risk events are occurring. Ultimately, it will reduce notably the overhead and many hours compiling data and organizing people.

Pharma R&D produces more and more research data in digital form. Data creation, processing, and analysis are often done by different people at different places and times because of factors like increasing automation and specialization. This talk will look at how to manage data for life science research in a way that can make the dependency between data producers and consumers transparent to enhance data-driven decision-making.

Scientists performing in vivo work within vivaria have historically been left on their own for many years. Oftentimes, outdated technologies such as pen and paper, mundane data entry, and even memorization have been used to document information throughout the experimental process. We will present the journey we took to enable scientists using 2D barcoded ear tags, thermal scanners, voice-enabled scientific LabFlows, as well as HoloLens.