Accelerating genomic workflows using NVIDIA Parabricks.

BACKGROUND: As genome sequencing becomes better integrated into scientific research, government policy, and personalized medicine, the primary challenge for researchers is shifting from generating raw data to analyzing these vast datasets. Although much work has been done to reduce compute times using various configurations of traditional CPU computing infrastructures, Graphics Processing Units (GPUs) offer opportunities to accelerate genomic workflows by orders of magnitude. Here we benchmark one GPU-accelerated software suite called NVIDIA Parabricks on Amazon Web Services (AWS), Google Cloud Platform (GCP), and an NVIDIA DGX cluster. We benchmarked six variant calling pipelines, including two germline callers (HaplotypeCaller and DeepVariant) and four somatic callers (Mutect2, Muse, LoFreq, SomaticSniper). RESULTS: We achieved up to 65 × acceleration with germline variant callers, bringing HaplotypeCaller runtimes down from 36 h to 33 min on AWS, 35 min on GCP, and 24 min on the NVIDIA DGX. Somatic callers exhibited more variation between the number of GPUs and computing platforms. On cloud platforms, GPU-accelerated germline callers resulted in cost savings compared with CPU runs, whereas some somatic callers were more expensive than CPU runs because their GPU acceleration was not sufficient to overcome the increased GPU cost. CONCLUSIONS: Germline variant callers scaled well with the number of GPUs across platforms, whereas somatic variant callers exhibited more variation in the number of GPUs with the fastest runtimes, suggesting that, at least with the version of Parabricks used here, these workflows are less GPU optimized and require benchmarking on the platform of choice before being deployed at production scales. Our study demonstrates that GPUs can be used to greatly accelerate genomic workflows, thus bringing closer to grasp urgent societal advances in the areas of biosurveillance and personalized medicine.

Are You Using the Right Electronic Health Record?

With the enactment of the Health Information Technology for Economic and Clinical Health (HITECH) Act in 2009, hospitals and physician practices across the country converted from a system of paper recordkeeping to fully integrated electronic health records (EHR).1, 2 With financial incentives in hand, there was a rush to market to acquire and implement these systems. Fast-forward 10 years, and it is apparent that the EHR space has significantly evolved in technology, processes, and policies.3 These changes should make organizations examine their EHR and organizational models and consider if they are using the best EHR to meet their organizational needs for the next 20 years. The National Institutes of Health (NIH) Clinical Center (CC) implemented its EHR in 2004 and, recognizing all of the new participants, technologies, and the advancement of clinical research needs since then, made the decision to embark on a comprehensive business case analysis to evaluate the best solution to meet the CC's and NIH's needs over the next 20 years. The goal was to answer this question: "Given the evolution of the EHR market, is the CC on the best platform to meet its needs now and in the future?"

Applications of artificial intelligence (AI) in ovarian cancer, pancreatic cancer, and image biomarker discovery.

BACKGROUND: Artificial intelligence (AI), including machine learning (ML) and deep learning, has the potential to revolutionize biomedical research. Defined as the ability to "mimic" human intelligence by machines executing trained algorithms, AI methods are deployed for biomarker discovery. OBJECTIVE: We detail the advancements and challenges in the use of AI for biomarker discovery in ovarian and pancreatic cancer. We also provide an overview of associated regulatory and ethical considerations. METHODS: We conducted a literature review using PubMed and Google Scholar to survey the published findings on the use of AI in ovarian cancer, pancreatic cancer, and cancer biomarkers. RESULTS: Most AI models associated with ovarian and pancreatic cancer have yet to be applied in clinical settings, and imaging data in many studies are not publicly available. Low disease prevalence and asymptomatic disease limits data availability required for AI models. The FDA has yet to qualify imaging biomarkers as effective diagnostic tools for these cancers. CONCLUSIONS: Challenges associated with data availability, quality, bias, as well as AI transparency and explainability, will likely persist. Explainable and trustworthy AI efforts will need to continue so that the research community can better understand and construct effective models for biomarker discovery in rare cancers.

FAIRshake: Toolkit to Evaluate the FAIRness of Research Digital Resources.

As more digital resources are produced by the research community, it is becoming increasingly important to harmonize and organize them for synergistic utilization. The findable, accessible, interoperable, and reusable (FAIR) guiding principles have prompted many stakeholders to consider strategies for tackling this challenge. The FAIRshake toolkit was developed to enable the establishment of community-driven FAIR metrics and rubrics paired with manual and automated FAIR assessments. FAIR assessments are visualized as an insignia that can be embedded within digital-resources-hosting websites. Using FAIRshake, a variety of biomedical digital resources were manually and automatically evaluated for their level of FAIRness.

Effect of battery longevity on costs and health outcomes associated with cardiac implantable electronic devices: a Markov model-based Monte Carlo simulation.

INTRODUCTION: The effects of device and patient characteristics on health and economic outcomes in patients with cardiac implantable electronic devices (CIEDs) are unclear. Modeling can estimate costs and outcomes for patients with CIEDs under a variety of scenarios, varying battery longevity, comorbidities, and care settings. The objective of this analysis was to compare changes in patient outcomes and payer costs attributable to increases in battery life of implantable cardiac defibrillators (ICDs) and cardiac resynchronization therapy defibrillators (CRT-D). METHODS AND RESULTS: We developed a Monte Carlo Markov model simulation to follow patients through primary implant, postoperative maintenance, generator replacement, and revision states. Patients were simulated in 3-month increments for 15 years or until death. Key variables included Charlson Comorbidity Index, CIED type, legacy versus extended battery longevity, mortality rates (procedure and all-cause), infection and non-infectious complication rates, and care settings. Costs included procedure-related (facility and professional), maintenance, and infections and non-infectious complications, all derived from Medicare data (2004-2014, 5% sample). Outcomes included counts of battery replacements, revisions, infections and non-infectious complications, and discounted (3%) costs and life years. An increase in battery longevity in ICDs yielded reductions in numbers of revisions (by 23%), battery changes (by 44%), infections (by 23%), non-infectious complications (by 10%), and total costs per patient (by 9%). Analogous reductions for CRT-Ds were 23% (revisions), 32% (battery changes), 22% (infections), 8% (complications), and 10% (costs). CONCLUSION: Based on modeling results, as battery longevity increases, patients experience fewer adverse outcomes and healthcare costs are reduced. Understanding the magnitude of the cost benefit of extended battery life can inform budgeting and planning decisions by healthcare providers and insurers.

Estimated hospital costs associated with preventable health care-associated infections if health care antiseptic products were unavailable.

OBJECTIVES: Health care-associated infections (HAIs) pose a significant health care and cost burden. This study estimates annual HAI hospital costs in the US avoided through use of health care antiseptics (health care personnel hand washes and rubs; surgical hand scrubs and rubs; patient preoperative and preinjection skin preparations). METHODS: A spreadsheet model was developed with base case inputs derived from the published literature, supplemented with assumptions when data were insufficient. Five HAIs of interest were identified: catheter-associated urinary tract infections, central line-associated bloodstream infections, gastrointestinal infections caused by Clostridium difficile, hospital- or ventilator-associated pneumonia, and surgical site infections. A national estimate of the annual potential lost benefits from elimination of these products is calculated based on the number of HAIs, the proportion of HAIs that are preventable, the proportion of preventable HAIs associated with health care antiseptics, and HAI hospital costs. The model is designed to be user friendly and to allow assumptions about prevention across all infections to vary or stay the same. Sensitivity analyses provide low- and high-end estimates of costs avoided. RESULTS: Low- and high-end estimates of national, annual HAIs in hospitals avoided through use of health care antiseptics are 12,100 and 223,000, respectively, with associated hospital costs avoided of US$142 million and US$4.25 billion, respectively. CONCLUSION: The model presents a novel approach to estimating the economic impact of health care antiseptic use for HAI avoidance, with the ability to vary model parameters to reflect specific scenarios. While not all HAIs are avoidable, removing or limiting access to an effective preventive tool would have a substantial impact on patient well-being and infection costs. HAI avoidance through use of health care antiseptics has a demonstrable and substantial impact on health care expenditures; the costs here are exclusive of administrative penalties or long-term outcomes for patients and caregivers such as lost productivity or indirect costs.

Characteristics and value of machine learning for imaging in high content screening.

Requirements for a flexible image analysis package for high content screening (HCS) are discussed. An overview of tools and techniques for image analysis and machine learning is given. Machine learning for classification and segmentation, the two fundamental elements of image analysis, is discussed. Next generation image analysis packages for HCS are reviewed. Recommendations for the development of image analysis solutions for advanced assays are given.