Data sharing among stakeholders in the development, access, and use of drug therapies is critical. In this talk, we describe our systems engineering approach with a methodology for developing a multi-stakeholder data sharing system, with its focus on high-level requirements that influence the design of the system architecture and technology choices. A realistic use case is addressed to illustrate the need for a scalable design for multi-stakeholder data sharing.  

High-performance computing technology helps life science pioneer new solutions to the complex computation problems. The diverse and dynamic computing demands challenge the agility of infrastructure capacity and capability, particularly when artificial intelligence continues to be adopted by the research field. There have been the arguments of where to adopt the next-generation HPC – through either on-premises or cloud computing. The concerns range from cost, data compliance, data gravity, workflow challenges, etc. Here presents an on-premises and public cloud hybrid HPC solution to converge the benefits of both sides and ensure business continuity.

With the advent of single cell sequencing at Pfizer, various departments began addressing the challenges of data management and processing in isolation. The newly formed Integrative Biology discipline made a strategic decision to develop a project gallery to share scRNAseq data across the organization. I will describe Pfizer’s partnership with Seven Bridges to design and deliver a powerful informatics platform that balances flexibility while fostering data discoverability, reproducibility, and reuse.

The combination of new therapeutic modalities, cloud computing, new scientific informatics vendors, analytics, and AI requires a service-based multi-vendor architecture to provide efficiency in the research process. Legacy systems duplicate many required scientific informatics functions within modality-specific applications. A utopian future of data-driven research requires compute and data systems to seamlessly interconnect to provide scientists the data and modeling results needed to make the best possible research decisions without compromise.

Over the last two years, HMS Research Computing has been working to define their Cloud Computing architecture and service offering. Join James Weatherell as he dives into the HMS Research Computing Cloud strategy. This presentation aims to highlight key challenges the HMS Research Computing team faced around Cloud adoption, along with a look at the cloud architecture and service offerings HMS Research Computing provides their researchers today.

Analytical methods are essential for pharma research. The primary purpose of analytical methods is answering (bio-)chemical questions, where the data generated is often used only once. Parallelly, we have data AI approaches that require consistent and suited data, which are not usually available. We present our journey on leveraging analytical data: data acquisition, capture, storage, conversion, and offering it back to scientists through a system built on the cloud environment following the FAIR principles.

Cloud-driven data analytics for transformation and innovation in healthcare: with continued growth and technology innovation in data analytics, cloud and AI - organizations are, more than ever focused on digital transformation. Big data, AI and cloud are enabling transformation through new products and services, helping organizations to become more data-driven at scale and speed of market. This session will focus on how to develop a successful digitization strategy with cloud-driven data analytics, innovation with AI and how to manage the challenges along the way with real life examples, use cases and experience from the industry.

Healthcare Data Ecosystem is multi-faced and developing a data ecosystem that can handle streamlined data ingestion, processing, and optimal consumption of this data in a self-service mode is both challenging and complex. This presentation gives a glimpse of a design pattern and approach to accomplish it in one of the leading Health Systems in New York. How the cloud infrastructure was leveraged for various tools and technologies and how the processes were set up around information lifecycle management to maximize access and visibility to data and reporting assets across the health system is presented.

Artificial intelligence and machine learning (AI/ML) are important tools for scientists and engineers to accelerate new discoveries. As in other areas of research, we have found that AI/ML uncovers new and previously unknown solutions to problems that have been overlooked. This talk explores a genomics pipeline with AI/ML workloads and Japan’s ABCI using Altair Grid Engine to manage AI/ML workloads and other batch workloads successfully.

In this Q&A-based session presented by RCH Solutions, hear well-known leader in the application of computing systems in biopharma organizations large and small, Clarence J. Wang, Ph.D., share lessons learned and best practices for the development of a Cloud-enabled, high-performing compute environment that supports the needs of both the business and IT, in organizations of all scale.

Bds is a simple yet powerful data analysis "orchestration" language. It is used to run data-intensive pipelines on tens of thousands of CPUs or hundreds of GPUs. Bds is a DSL (Domain-Specific Language) that allows you to quickly develop data analysis pipelines on your laptop and then run them on large servers, cloud, or High-Performance Computing infrastructure.

Bioinformatics researchers often find their projects limited by the time to perform alignments. ElasticBLAST solves this problem, able to BLAST hundreds of thousands of DNA or protein sequences by breaking the task into batches and distributing those to multiple cloud instances, using cloud native scheduling systems. This allows the researcher to move ahead with their projects more quickly. ElasticBLAST can search a user provided database or a popular database hosted on the cloud by the National Library of Medicine. We describe the architecture of ElasticBLAST, show examples of the work it can do, and discuss related cloud resources.

To fully benefit from exponentially growing biomedical data, researchers need easy access and computation capabilities for big data. In our efforts towards laying the foundation for a biomedical dataverse, NLM’s National Center for Biotechnology Information (NCBI) in partnership with NIH STRIDES has made the entire Sequence Read Archive and computational tools accessible on the cloud to create an equitable ecosystem where NIH-funded data is FAIR.

You're a research organization and all of a sudden some requirement comes down for you to do some robust security for the software you build and operate. How do you get there? What parts are security and what parts are compliance and what "really" counts?

Hybrid multi-cloud is becoming the norm. For their advantages, hybrid multi-cloud environments introduce several data management challenges and requirements. BIO-IT organizations deal with hybrid multi-cloud complexity, with the need to facilitate data collaboration across many research entities, and with data governance, compliance, and privacy requirements. This session will overview the state of the market for multi-cloud data management tools, and considerations for addressing the most pressing related pain points.

With the increasing demands of biomedical research for the ongoing COVID-19 pandemic, a secured computing infrastructure is needed to support the informatics efforts. The Precision Vaccines Program Data Management Core at Boston Children’s Hospital designed a digital infrastructure using Amazon Web Services cloud computing capabilities to facilitate data management, data storage, and data analysis of study endpoint and safety data allowing collaboration across biomedical centers in the United States.