Biosurveillance Signatures for and public health

A Los Alamos National Laboratory Implementation Plan for the Science of Signatures Biological Signatures for National Security Los Alamos National Laboratory

Biosurveillance at Los Alamos Table of Contents

Los Alamos National Laboratory’s charge is to develop science and technology that will make the nation safer and enhance our global standing. This breadth of mission scope requires careful Biosurveillance at Los Alamos ...... 2 internal planning and effective cooperation with external partners and other governmental agencies. The document you A National Imperative ...... 4 are holding is one of the products of ongoing planning efforts that are designed to bring to bear the Laboratories unique Laboratory Planning...... 5 capabilities on problems of the greatest significance.

To those unfamiliar with the extent of our science, it may seem odd that Who should read this...... 5 our planning includes such a strong biology focus, yet our work in this area extends all the way back to the Manhattan Project and the birth of large scale LANL Strategic Context...... 6 government-sponsored research. Following World War II, the Laboratory began programs in health and radiation physics that expanded to become the robust Biosurveillance Overview...... 8 bioscience program we have today. Along the way, we have made contributions such as the Human Genome Project, the invention of flow cytometry, HIV Case Studies: Tuberculosis and Bird Flu...... 10 virus management, and more. Our greatest advances come at the interface of biology and high performance computing, modeling, the physical sciences, and Goal 1: Diagnostics...... 12 engineering.

Our objective now is to look to the future and apply our strengths in these areas Goal 2: Predictive Modeling...... 14 to new problems in biosurveillance. The challenge is not small; we must manage and marshal our internal resources, select the correct problems to which those Goal 3: Complex Data Analysis...... 16 resources should be applied, and then coordinate with external collaborators in academia, government, and industry. Next Steps...... 18

We intend this high-level plan to help us both communicate internally and initiate conversations with the Departments of Health and Human Services, Defense, Security, and others into how we might apply our expertise to helping solve biosecurity problems.

Additional resources can be requested from the contacts listed on page 18. Strategic planning is a living process, and we welcome your feedback and questions. LANL conferences in biosurveillance. 2006 2007 2010 2011 Nancy Sauer is the Associate Director for Chemistry, Life, and Earth Sciences, which oversees the Science of Signatures effort for the Laboratory. She can be reached at (505) 606-2266.

Biosurveillance at Los Alamos is jointly managed by the Science of Signatures and the Information Science and Technology (IS&T) pillars (see p6). Capabilities in supercomputing, large scale data analysis, and modeling combine in a unique way with our more conventional bioscience capabilities to address issues in disease surveillance. Shown here are the supercomputers Moonlight and Cielo. The The Laboratory has been involved in organizing national and international biosurveillance technology Cielo system is a classified supercomputer advancement conferences since 2006. capable of analyzing the most sensitive challenges in biosecurity. Biological Signatures for National Security Los Alamos National Laboratory

A National Imperative Laboratory Planning investments in hiring, facilities, equipment, as well as additional planning. There is a recognized national need for Strategic planning necessarily cascades through advanced biological signature technologies an organization, and this biosurveillance Our stakeholders: It is intended as a that could, among other things, address overview document grew out of a lengthy communications document for those outside disease surveillance. The Laboratory’s strategy planning process that involved both internal the Laboratory who have current or potential in biological signatures maps directly to the and external workshops and reviews. See the programs that might benefit from the expertise national posture in . In 2014, the organizational chart in the next section for a available at Los Alamos. With it, we hope to government is projected to spend almost $6.7 sampling of the organizations that contributed. begin a dialog that will help us understand the billion on biodefense, with the bulk of that technological and scientific requirements of targeted to programs having application in Beyond this biosurveillance planning, the our customers and to inform them of where we nondefense space as well. See figure, right. Laboratory is engaged in numerous other might be of assistance. strategic exercises. Collectively their intent is The intent of this funding, which has grown to coordinate and marshal people, equipment, Our collaborators: Biosurveillance is a cross- each year from 2012 to 2014, is preparedness, and facilities behind important technological organizational and multidisciplinary endeavor. prevention, and mitigation of biological threats problems within the scope of our mission. To be successful, we must collaborate with to civilians. Each of these areas has a significant, other science and technology organizations, even primary, signature related aspect. Civilian biodefense budget by agency, FY2014 (in Who Should Read This the public health sector, local and national $millions). Federal Agency Biodefense Funding, governmental bodies, and through them with The goal of the Laboratory is to partner with FY2013-FY2014. Tara Kirk Sell, Matthew Watson. This document was written with multiple the civilian population that we seek ultimately the agencies that are executing our national Biosecurity and : Biodefense Strategy, audiences in mind. to serve. This document explains what we hope biodefense strategy and provide science and Practice, and Science. September 2013, 11(3): to achieve and where we will require partners to technology solutions to biological threats using 196-216. Our staff: This document is the high level succeed. the unique resources of the Laboratory. Laboratory strategy for biosurveillance. It is intended to unite Laboratory scientists behind a single strategic direction and to the guide the research and efforts of our staff. It will inform The Laboratory’s approach is informed and guided by both the recently-published National Strategy for Biosurveillance (White House), and a study by the National Academy of Sciences that focuses on the importance of cooperative engagement amongst the The National Security Sciences building nation’s science and technology leaders. from the Los Alamos townsite.

The Laboratory provides technology, expertise, and on-the-ground support to the agencies tasked with managing biosurveillance threats. We do so in collaboration with a wide variety of partners and collaborators.

4 5 Biological Signatures for National Security Los Alamos National Laboratory

LANL Strategic Context under the Science of Signatures. However, SoS Scientific Approach there is a significant component of both IS&T • Nuclear and Radiological The science pillar concept is a primary tool and Materials required in biosurveillance. Our scientific strategy is to discover new • Chemical and Materials the Laboratory uses to plan how we will Laboratory planning efforts in biosurveillance signatures, revolutionize measurement of • Biological accomplish current and future missions, have therefore included leadership from these signatures, and deploy new technologies in the • Energy including the biosurveillance mission that is the pillars as well. field. Each of these three components has • Climate focus of this document. There are four science distinct characteristics. • Space pillars: 1-Materials for the Future (Materials), Science of Signatures Discover signatures: Identify signatures of 2- Integrating Information Science and The other five areas of leadership are developing chemical, biological, radiological, nuclear, Technology for Prediction (IS&T), 3-Nuclear The Science of Signatures strategic plan (pub- strategic plans similar to this biosurveillance and explosives threats and of climate, energy and Particle Futures (NPF), and 4-the Science lished 2012) was chartered by the Principal plan. Each individual plan will nest with the and health security impacts. In essence, of Signatures (SoS). Each of the pillars has Associate Directorate for Science, Technology, larger SoS plan. signature discovery is determining those discrete science goals that are fundamental to and Engineering (PADSTE) and written by a measurable phenomena that uniquely SoS is tightly integrated with the other pillars building the Laboratory’s future science and team of more than 35 managers and scientists identify and characterize properties within along the themes of Discover, Revolutionize, technology base. These pillars support each from across the Laboratory following a two complex environments. Deploy. other, and interfaces among the four pillars are year process. Its grounding principle is that we leveraged for the benefit of all four. must be able to identify and characterize threats Revolutionize measurements: For threat- before we can understand them or take action specific signatures, develop entirely new The fundamental precept of this approach is to mitigate them. Signatures are the unique measurement technologies, methodologies, that the greatest science breakthroughs will The Science of Signatures Pillar is elements that allow us to locate threats within or strategies or develop transformational come as we approach difficult problems in their environments and describe them. advances in the current state-of-the-art. revolutionary ways. This multidisciplinary • An Institutional effort led by the In essence, how can sensitive and specific approach draws upon physicists, materials The Science of Signatures pillar addresses Associate Directorate for Chemistry, measurements be made in entirely new scientists, chemists, computer scientists, emerging challenges by developing science Life, and Earth Sciences. ways and/or how can new phenomena theoreticians, biologists, earth scientists, space and technology to detect these threats. Our (signatures) be measured? scientists, engineers, mathematicians, and complete technological toolbox is applied to • An organizing principle built on numerous other disciplines to solve important signature science from across our mission areas Forward technology deployment: historical strengths and technical national security science problems. The pillars of global security; nuclear defense; and energy, Make measurement technologies leadership. approach gives these experts a framework for climate, and health. Critical components are and methodologies practical through working together and allows them to apply their the discovery and detection of signatures to engineering. This includes prototyping • One of four science plans to shape skills across the traditional boundaries of their enable understanding of component species or of sensors and instruments for field and manage the future of science and disciplines. processes that have a major impact on a large, deployment and systems integration of technology at the Laboratory. complex system. sensor networks. In essence, how do we The science pillars also inform our investments bring science advances to the real world in science and engineering, guide recruitment Specifically, we characterize measures, signals, • Supported and mandated by DOE. in a way that provides feedback into and training strategies, and serve as a and properties in or of complex systems to signature discovery and/or revolutionary framework for our partnerships with other order to detect or attribute change; predict • Important to our immediate measurement technologies? leading research institutions worldwide. systems behavior across scales in space stakeholders and the nation. Effective biosurveillance incorporates a wide (molecular to global) and time (near-term to The strategy ofDiscover, Revolutionize, variety of signatures and thus ownership of geologic), and assess impacts to the system of Deploy is applied to the Los Alamos SoS • Reshaping the way we think of the associated strategic planning efforts fall change. areas of scientific leadership: multidisciplinary science. From concept to application: this hand-held “dipstick” was developed by LANL scientists to detect multiple pathogens quickly Biological approaches to detecting pathogens, diagnosing and and with minimal resources. treating disease, and predicting spread are useful for mitigating all The scientists who invented types of biothreats. The impacts of bioterrorism and pandemics are it are now pursuing its unknown but can potentially be great when considering the social, development through the political and economic effects that inevitably follow. By comparison, startup company Mesa annual deaths from chronic and infectious disease are comparatively Tech International. predictable in both impact and likelihood but are still very destructive. A comprehensive plan for biothreat reduction must consider both. Donald Montoya, graphic.

6 7 Biological Signatures for National Security Los Alamos National Laboratory

Biosurveillance Overview To succeed, the national effort will require new signatures and genomic sequencing, also areas LANL capabilities not previously used in tools and approaches for in which Los Alamos has made significant biological research show potential for our Biosurveillance is the collection, analysis and • Point-of-care diagnostics, contributions. evolving biosurveillance effort interpretation of data to help monitor for • Predictive modeling, and he presence of pathogens in plants, animals, • Decision support through complex data Theoretical biologists and epidemiologists Los Alamos has a rich legacy of engineering humans, food, and the environment. Its goal analysis. at Los Alamos are widely recognized for capabilities that have traditionally been used is to save lives by informing intervention their development of agent-based, predictive to build spaced-based instrumentation and strategies and by guiding public health These three goals (which are detailed in the models that can reveal the potential path of sensors as well as remote and unattended decisions. Biosurveillance following pages) play to the an outbreak, as well as the effect of different sensors. These capabilities have in the past and is the central theme of Our vision is for the Laboratory Laboratory’s current and mitigation strategies. This work leverages could in the future be used to adapt and deploy biosecurity. to integrate and use unique historical strengths and a strong Laboratory high-performance detection and sensing discoveries made on the strengths in biological, physical, thus comprise the structure computing program that includes some of the bench to real-world applications in the field and Historically, biosurveillance and information sciences to of our biosurveillance fastest super computers in the world. in clinics. has been dominated by the safeguard the nation through strategic plan. An integrated military’s concern for threat advances in biosurveillance. approach allows the Laboratory capabilities in Information Science Finally, image analysis and machine learning pathogens (commonly called Laboratory to draw upon and Technology (e.g., computational co-design, capabilities that have been developed at the “select agents”), but more recently its scope has the multidisciplinary science that has been our data-science-at-scale, analysis of complex Laboratory to extract knowledge from high broadened to include identification of emerging hallmark for more than 70 years. networks) have also been used to address “big content but poor resolution images. High infectious disease and a concern for the civilian data” problems that range from climate model- volume sensing data are being explored as populations. An additional permutation is a For instance, Los Alamos experimental ing to cosmology and are characterized by the a means to extract diagnostic information concern for safeguarding US military personnel biologists specialize in novel in vitro methods to integration of diverse and heterogeneous data from image-based sensors. These emergent in regions where infectious disease coexists generate reproducible, high-quality antibodies streams with quantified uncertainties. capabilities, as well as those already being used with diseases caused by threat pathogens. The and other affinity reagents such as peptides. to address biosurveillance, form the bedrock of ultimate goal of the evolving national effort This approach generates better results and These strategies are now being used to address our approach. As shown in the organizational in biosurveillance is an intelligent, real-time avoids the use of animals in research. These the formidable data integration and analysis chart below, this strategic plan requires that system where differing surveillance data capabilities, when combined with other LANL problem required by an intelligent, real-time expertise be drawn from across the Laboratory. streams, which range from social media to expertise in biomarker discovery and assay biosurveillance system. Additionally, other clinical and point-of-care diagnostics, can be development, have extensive applications for integrated and rigorously analyzed to provide detection and diagnostics. Other strategies guidance to decision makers from the local to for diagnostics include nucleic acid-based Science, Technology & Engineering Global Security Weapons the national level. Programs Chemistry, Engineering Experimental Information Theory, Threat Life, & Earth Sciences Physical Technology Simulation, & Identification Plutonium Biosurveillance components: early warning, early detection, Sciences Sciences Computation & Response Science & situational awareness, and consequence management. Manufacturing ADCLES ADE ADEPS ADIT ADTSC ADTIR Accelerator Materials Departmental Computer, Weapons Bioscience Operations & Physics & Computing Computational,& Defense Systems and Analysis Technology Applications Services Statistical Sciences Engineering Ecosystem Potential cases prevented/ & Experiments baseline Rapid international spread prevented Applied Materials Network & High responsep Early Chemistry Engineering & Science & Infrastructure Performance Climatology Technology Technology Engineering Computing Intelligence, Analysis, and Technology Weapons reporting analysis Physics Earth & Prototype Los Alamos Software & Vector Environmental Neutron Science Applications Theoretical Fabrication ecology Sciences Center Engineering Intelligence and Space Research Animal disease Physics Nuclear Engineering & Nonproliferation Number of new cases

Science Program National Security Science Resource - X--20 1 1 4 7 10 13 16 19 22 25 28 31 34 37 40 Offices Education Center Office Days Strategic Outcomes Office Risk Detection Sensitive & prediction Risk Office of Nuclear Diagnostics Special Program Threat Counterintelligence Emerging Threats Nonproliferation alert Operations & Security alert Disease models

Medical countermeasures Global Security Intelligence, Technology Program Office Defense & Transfer This timeline of events in a hypothetical disease outbreak shows that detection and diagnosis in the Counterterrorism first week can enable a rapid response that would significantly reduce the overall spread of the disease. Laboratory organizational chart showing in blue the organizations that currently contribute to biosurveillance However, looking backwards in time, there are opportunities to predict coming diseases prior to the first science development. The Laboratory as a whole has a budget of over $2B and more than 9000 employees, reported case based on broad categories of information such as ecosystem and climate data, vector ecology and while our “bench strength” is already great, it can clearly go much deeper should the science challenges data (e.g., mosquito populations), and diseases in animal populations. A comprehensive biosurveillance demand it. strategy spans the spectrum. Leslie Sandoval, graphic.

8 9 Biological Signatures for National Security Los Alamos National Laboratory

Tuberculosis Outbreak Kwazulu Natal, South Africa Bird Flu Outbreak Case Study Case Study

In 2004 in the small village of Tugela Ferry in Kwazulu Natal, South Africa, In April, 2009, a 10-year-old boy in San Diego, California, was diagnosed 54 people tested positive for Mycobacterium tuberculosis (TB). In 16 days, with a novel strain of H1N1 flu. Newspaper headlines asked if it was a 53 had died. The rapid progression to death in this population caught the bioterror attack. The query arose again in 2011, when a highly unusual world by surprise. TB is generally thought to be a chronic, slow moving strain of E. coli spread rapidly through Germany. bacterial infection that is readily treatable with a cure rate of nearly 100%. But the Tugela Ferry strain was different, partly because the infected Every time a novel pathogen appears—on average, one every year—the individuals were HIV positive, but mostly because this was a new mutant “what if” question arises. What if that flu strain had instead been a deadly strain of TB. It was resistant to not only first line drugs but also second line H5N1 avian flu engineered by terrorists? Release of a high-fatality H5N1 drugs developed recently and used to treat the multi-drug resistant strains A Path Forward: Los Alamos National Laboratory can serve as strain able to spread rapidly among humans could result in millions of (MDR/TB) that began appearing in the 1990s. a science and knowledge resource for the agencies and entities deaths. Clinical diagnosis of a terrorist attack would be too late for optimum that are the first line of response. The three part strategy outlined containment. The most effective way to reduce biological threats to national In less than a decade, this extensively drug resistant TB (XDR/TB) has in this brochure can be a major force in mitigating biological security is to avert them early. A major focus of LANL science is the spread throughout South Africa and in recent years has been found in threats such as the XDR/TB outbreak, a bird flu outbreak, or an identification of threat signatures and technology development for early many other countries. One of the world’s leading experts on this outbreak emerging disease that makes a species jump. detection—before anyone gets sick. (Willem Sturm of the University of Kwazulu Natal and then Director of the Kwazulu Natal Research Institute for Tuberculosis and HIV) believes In 2012, the international community panicked when US and Dutch the TB outbreak could have been stopped at Tugela Ferry with a Prompt diagnostics: Multiple efforts at scientists evolved potentially human-transmissible strains of H5N1 bird comprehensive biosurveillance strategy similar to what is proposed. LANL are improving detection through flu, which has a 60% mortality rate in humans. Concern focused on the field-ready diagnostic tools such as the potential for terrorists to misuse the research as a recipe for a bioweapon. Diseases such as this will continue to emerge and we must prepare. optical biosensor that could quickly and Los Alamos is developing capabilities to counter this type of emerging An effective and comprehensive biosurveillance strategy has a flexible more reliably detect active infections: threat by integrating information and technologies from a variety of sources. approach that can be implemented in different scenarios: identify and e.g., XDR/TB or H1N1. They seek to disrupt early phases of threat development in various ways, stop the disease at inception; contain its spread; inform local, national, including enhanced export control, which attempts to reduce the spread and global strategies for management. In Tugela Ferry, better diagnostics Data integration and analysis: of enabling technology. LANL is also developing capabilities to enhance would have made a big difference; authorities needed to know who had Integration of disparate data sources for detection of technical expertise recruitment and improving tools to active infection. Most current TB tests are fraught with issues—some can analysis, including social media such as detect the acquisition of reagents and equipment for nefarious biological show false positives in those with a prior infection or a vaccination, while Twitter, can provide valuable information laboratories. others can show false negatives in certain populations (those with HIV, for on emerging disease hot spots in real example). Moreover the tests must be feasible in resource poor settings time. If terrorists were to succeed in producing a human transmissible strain, such as Tugela Ferry. poor containment during development could result in a local outbreak of Decision support through complex highly-fatal influenza, which would produce threat signatures for detection The rapid and accurate field identification of active infection allows data analysis: Epidemiological models at Poor containment during development could result in a local outbreak immediate initiation of treatment and quarantine protocols to minimize LANL have been used in past epidemics of highly-fatal influenza, which could produce other threat signatures for further spread. It also provides a filter to triage patients for more lab- to predict disease progression and to help detection. To improve early detection of anomalous events like these, LANL intensive strain identification. In a few hours, laboratory methods such identify measures that could be taken to is contributing to efforts led by DoD, the State Department, and DOE to as PCR can identify strains which in turn generates a recommended stop it. Other models are used to help enhance disease surveillance capabilities in central Asia, the Middle East, treatment protocol, indicates prevalence of drug resistant infections, and identify potential outbreaks using animal and Africa. describes transmissibility. Sequencing technologies provide the next disease models and climate data. level of characterization and facilitate the identification of new strains and Finally, if the enhanced strain was released, CDC and DHS would recruit a previously unreported mechanisms of drug-resistance. set of epidemiologists to forecast the progression of the epidemic in order to assess the possibilities for containment. LANL continues to build on its This diagnostic approach provides a reliable data stream for inclusion with history of producing cutting-edge disease forecasting by developing real- many other discordant data streams (e.g., syndromic and social media time tools that will incorporate human behavior patterns data) that can then be used to track disease progression at the population (for example seclusion at home or urgent flight to other level for epidemiological forecasting. The data integration and fusion locations), which are likely to deviate from combined with predictive modeling can normal in such an emergency. Forecasting form the “data to knowledge” disease progression based on real-time biosurveillance system that can be conditions would be vital in such a scenario used to help guide decision to enable the US government to identify the makers at all levels. In the case best options for disease mitigation. of the Tugela Ferry XDR/TB outbreak, it is this information that could have been used to contain the spread of this Air travel map. For biosecurity in the 21st Century, serious, high mortality there is no such thing as a “local threat.” A obscure disease that now threatens disease in a small African village is of immediate not only South Africa but also importance to the US, as is the development of new the entire world. biothreat agents. WikiCommons 10 11 Biological Signatures for National Security Los Alamos National Laboratory

Goal 1: Diagnostics Drug resistance is increasingly a problem in Goal 1.1: Create universal approaches for the both bacterial and viral pathogens, and often discovery and detection of all pathogens by Early detection of biothreats is critical the mechanism of resistance is unknown. for controlling their spread. Clues for an Understanding anti-microbial resistance is • The discovery of pathogen biomarkers impending outbreak are varied and require critical to effective countermeasures. Further, directly from infected hosts. creative, orthogonal approaches such as patients in regions like East Africa manifest • Development of new strategies for early monitoring animal populations and climate with multiple health issues which can detection/diagnosis of infection. data to find emerging pathogens, as well as complicate diagnosis: co-infection with other communication data to intercept terrorist plots. pathogens (e.g., tuberculosis), species jumps • Development of approaches for from wild animals to humans, vector-borne discriminative detection: early detection Timely detection of disease that is already active diseases, simultaneous infection from multiple should be accompanied by discriminative in the human population is often constrained pathogens, and myriad chronic infections and diagnosis to facilitate effective Los Alamos computational biologists develop and by the limits of diagnostic tests. There is a existing co-morbidities (HIV, malnutrition, countermeasures. validate genetic markers for high-consequence broadly acknowledged need for better point- diabetes, and others). pathogens for potential use by DHS BioWatch and of-care (POC) diagnostics. However, to realize • Innovation for accessory tools and reagents, CDC Laboratory Response Network labs. Pictured a future goal of reliable POC diagnostics for all Understanding of the above factors is critical including recognition ligands, for effective below is a BioWatch detector in front of the Capitol infectious diseases, known and unknown, it is to achieving real-time global biosurveillance, biodetection. Building in D.C. critical to develop a new toolset for pathogen which integrates effective forecasting of identification and characterization. outbreaks using diagnostics and modeling. Goal 1.2: Improve strain identification for characterization of drug resistance and LANL’s technology roadmap for Accordingly, the LANL science community related virulence factors by biosurveillance includes this critical is focused on diagnostics for effective requirement for thorough characterization biosurveillance. These encompass field-ready • Advancing phenotypic vs. genotypic of pathogens at both regional and advanced diagnostics for the identification of active detection. Linking phenotypic antibiotic research laboratories. Pathogen characterization infection (for uses as varied as soldiers in resistance (MIC levels) with genotypic is critical to the identification of new strains remote field locations to home-based diagnostic variation will reveal the genetic mechanisms and serovars, predicting pathogen evolution kits), to more involved/comprehensive assays responsible for the various levels of and transmissibility, and tracking emerging for strain identification and characterization. antibiotic resistance. drug resistance. All of these factors are critical There are at least three critical technological for determining the path forward in the design approaches that can help us realize this goal: • Detecting/characterizing known vs. novel of vaccines and countermeasures, and are also antibiotic resistance mechanisms. significant for real-time biosurveillance through 1) Fieldable/rapid diagnostics for immediate • Creating culture-free strategies for genetic predictive epidemiology. pathogen identification, This rapid ultra-sensitive biosensor is a multiplex, analysis. multi-channel detection system that uses photo- 2) Nucleic-acid-based assays for stable and tunable quantum dots. Optimized for identification of known strains and markers • Resolving the sequencing challenge vs. field-based and environmental detection, it has of drug resistance, informatics challenge. been validated for the detection of cholera toxin, influenza, anthrax, tuberculosis, breast cancer, 3) Genome sequencing (DNA and RNA) Goal 1.3: Develop smart sensor interfaces adapt- E.coli, and other agents using single mode for comprehensive identification of all able to any transduction system to facilitate waveguides pathogens and novel markers of drug- deployment. resistance. • Detection systems and sensor prototyping Each of these technical approaches offers for POC diagnostic applications in complimentary information that together can biosurveillance. help in the more efficient identification of the • Data fusion and extraction. pathogen, minimize redundancy and decrease wasted effort. Together, these technologies allow Right: Our ultimate goal is a universal approach Harshini Mukundan (left) watches as fellow Los us to address the following strategic objectives that detetcs all pathogens. One possibility is Alamos bioscientist Elizabeth Hong-Geller loads for efficient diagnostics. functional validation of sRNAs in virulence a vertical electrophoresis gel, used to separate (pathogen) or immune response (host). A basic proteins and small molecules. This technique is understanding of organism interactions can useful for developing improved methods to detect lead to ways to detect infections, potentially pathogens. prior to manifestation of symptoms. Shown here is a high content cellular analysis.

12 13 Biological Signatures for National Security Los Alamos National Laboratory

Goal 2: Predictive The models developed at LANL can be Based on these capabilities Los Alamos has combined for use in a variety of situations. developed the following strategic objectives in Modeling Effective epidemiological models in Predictive Modeling for the next five years: biosurveillance require appropriate prevalence LANL scientists integrate predictive models Goal 2.1: Build epidemiological models capable with unique data streams and assays to address data and an understanding of pathogen transmission, risk factors influencing outcome, of incorporating disease mechanisms from biosurveillance needs. These models allow diverse data sets. the Lab to collaboratively solve related public and quantification of the effectiveness of mitigation strategies. In some instances, such health problems in order to both cost-share • Assess the risk of Staphylococcal and as in the 2002-03 Severe Acute Respiratory and greatly improve the fidelity with which the Salmonella co-infection in pediatric Syndrome (SARS) outbreak, analysis of high national security mission is carried out. malaria patients. quality data has allowed effective control of an Predictive modeling capabilities have already epidemic. • Assess the risk of co-infection from HIV been applied to complex health problems A model of a multi-drug efflux pump in Burkholderia; and TB. proteins such as this one are a leading cause of (such as methicillin-resistant Staphylococcus In other instances, such as the 2009 pandemic influenza, the role of epidemic forecasting can multiple-antibiotic resistance. A LANL interdisciplinary aureus), anticipating and preparing for a • Analyze the impact of vector-borne team seeks to understand how such genes achieve naturally occurring epidemic (e.g., influenza), be used to determine the allocation of scarce pathogens in biothreat scenarios. substrate specificity and how their expression and or assessing the risks from or responding to vaccines and antivirals, identify risk factors of activity are regulated so that they can be effectively targeted with a new generation of antibiotics. bioterrorism (e.g., engineered pathogens). poor outcome, and optimize case definitions for Goal 2.2: Develop theory and modeling to predict Models can fill in incomplete information, patient care and implementation of disruptive disease progression in a host. enable uncertainty quantification, and guide public health measures. In this type of situation, an important attribute of a successful predictive • Develop a link between pathogen levels, information collection. They fall into three target cell death, immune response and general categories. model is its ability to effectively communicate likely scenarios, potential high-consequence disease symptoms in a host. scenarios, potential “windows of opportunity” 1. Epidemic simulation can project outcomes • Predict how the metabolic state of for control measures, and situational awareness by using existing infection rates and the target cell is correlated with virus of sufficient detail to communicate risk varying assumptions about transmission production. mechanisms, behaviors, and control effectively to the public and implement control strategies with a goal of informing and measures. • Model the effect of drug resistance and optimizing decision-making during an emergence of a new pathogen in disease Finally, when integrated these three types outbreak. progression in a host. of predictive models can help anticipate the 2. Evolutionary analysis predicts how emergence of novel pathogens—as well as to Goal 3.3: Provide decision makers with validated diseases will change. When performed on identify the sample streams from across the epidemiological models for disease forecast and appropriate sets of pathogen samples it can globe required to address the problem. prediction. identify signatures to guide development and use of diagnostics and medical The Lab’s capabilities that enable predictive • Benchmarking of epidemiological models. countermeasures. modeling include: • Application of validated epidemiological 3. Disease models relating molecular • Agent (subject) based models/discrete models. Evolutionary models can be applied to genomic data to mechanisms to virulence, drug resistance, event models. understand global drivers of epidemic spread, identify functional markers, optimize information content from and immune response can provide • Analytical models. diagnostics, and guide sample collection, as in this important insights when potentially novel example from Ford, et al. published in Tuberculosis in pathogens are observed and can guide the • Host response modeling. 2012. development and interpretation of the numerous potential sources of clinical, • Genomic signatures for host/pathogen Right: Following the “Amerithrax” attack in 2001, Los Alamos scientists began optimizing methods of detecting laboratory, and non-medical information characterization. anthrax and modeling risks. At the time, public health when computing prevalence, risk factors, authorities grappled with the idea of a widespread crisis: and efficacy of control measures. • Modeling zoonotic reservoirs. how many letters would there be, what are the risk factors for various populations, should therapeutics be • Uncertainty quantification (UQ). disseminated or are the repercussions too great? Many of these questions could now be addressed using LANL modeling and detection techniques.

14 15 Biological Signatures for National Security Los Alamos National Laboratory

Goal 3: Complex Data capture widely varying data sets. All of these capabilities will be leveraged to address the considerations must be taken into account for sustainability challenges for developed tools and Analysis effective decision support through complex data resources and add functionality. analysis. This goal leverages advances in information Goal 3.1: Develop and assess new data streams science and technology (IS&T) to help public Technical challenges: that can be used for disease surveillance. health, agricultural, and national security • Real data streams are messy, vary in quality, • Obtain reliable data streams and data sets, agencies make timely decisions during outbreak and may be biased or incomplete. and assess content and quality. and biothreat scenarios. Doing so requires • Relevant data are complex, highly dimensional, • Develop new data streams and characterize collection, processing, integration, and analysis and have widely different structures, possibly them in the context of biosurveillance of multiple data streams and the development requiring aggregation and/or disaggregation decision-making. Complex data analysis of data analytics tools to transform data into • Many relevant data streams must be analyzed requires a combination of hypotheses and A suite of tools is being developed at LANL to actionable information. This information enhance situational awareness of an ongoing locally rather than ported to a central site discovery-driven approaches. can then be used by federal agencies, the event and transform data into actionable for analysis by virtue of privacy concerns. • Develop an evaluation framework to assess private sector, international, state, local, information. The BRD is a resource for validating This requires computational co-design of both traditional and non-traditional data tribal governments, and nongovernmental appropriate hardware, software, modeling, and information about disease outbreaks. The streams. BaRD is a database that catalogs and classifies organizations. data analysis tools that anticipate analyses for epidemiological model-specific information. The local sites (the small) that will scale to all sites Complex Analysis of Big Data: 21st century Goal 3.2: Develop and integrate LANL SWAP is an app to provide a context for a rapidly (the big). biosurveillance involves complex analysis of capabilities in disease biology, public unfolding event through graphical visualization. • Data volume may be large and continuously “Big Data,” which requires integration, analysis, health, medical diagnostics, statistical and generated, exceeding the rate (velocity) at and interpretation of large disparate data computational analysis, and computer science which they can be analyzed or stored. gathered from multiple sources. These sources to create complex data analysis and decision • There are limits to our ability to formulate include clinical diagnosis, syndromic analysis, support tools. complex queries and convey the associated ‘omics studies on clinical samples, biomedical uncertainty inherent in the analysis process. research literature, electronic health records, • Identify data streams that can be used synergistically. climate/geographical surveys, lab, imaging, Our multi-disciplinary capabilities address these • Develop robust integration algorithms for and pharmacy systems, infrastructure systems, challenges and achieve efficient data driven disparate data streams that can be used for on-site and remote patient monitoring, email, decision support through complex data analyses. and social media. Analysis requires tools and disease surveillance. techniques for acquiring, cleaning, and exploiting Transforming data into actionable information: • Develop robust statistical models and data data using statistical techniques. The results Situational awareness is important for both visualization models that process complex enable situational awareness and provide robust early warning and early detection of outbreaks. and diverse data streams and support answers to questions of interest. Analytics and tools that furnish information situational awareness. • Build computational infrastructure to on how an infectious outbreak would either PARAVIEW is an open source visualization tool As in other applications (climate, cosmology etc.) emerge or unfold enhance situational awareness support efficient storage, access, and real- Big Data in biosurveillance has the attributes of time distributed processing and analysis of developed for the weapons program that has for decision makers/analysts/public health successfully been used to visualize output from four V’s: volume, velocity, veracity and variety. biosurveillance data. officials and support planning for prevention or large-scale simulations, including Episims. “Volume” refers to the size of the data set. High mitigation. • Develop tools and apps that enhance definition spatiotemporal video images or ‘omics situational awareness in an ongoing event of data can quickly fill up multi-terabyte disks. Data sharing and expert analysis of incoming infectious disease spread. However, it is not just the sheer data size that information are equally critical. LANL is must be managed; the “velocity” with which the developing new capabilities for decision support EpiSimS-computed attack rate by census tract data is accumulated is also an important factor. It in infectious disease surveillance. Robust and for a 1918-like pandemic influenza in southern might come from Twitter, Facebook, or sensors. comprehensive frameworks for data stream California. EpiSimS is a stochastic, agent-based and model characterization underlie these simulation that models the spread of disease in regions, allowing for the assessment of disease To be effective, it is necessary to rapidly scan, tools – they describe fundamental elements of prevention, intervention, and response strategies. store, process, and analyze this data. Additionally, infectious disease surveillance. The resulting data It explicitly represents the daily movements and many biosurveillance data streams are biased and transformation tools are intended to be accessible incomplete and require up-front data assessment interactions of synthetic individuals in a city or to the global biosurveillance community. We region, including their interactions with others. It and cleaning as well as error quantification will build new partnerships and collaborations is used as an experimental test bed for analyzing to address trustworthiness (or “veracity”). through subject matter expert panels developed the consequences, feasibility, and effectiveness Finally, “Big Data” contains a great “variety.” in conjunction with these tools. LANL IS&T of response options to disease outbreaks. They are collected from multiple sources and

16 17 Biological Signatures for National Security Los Alamos National Laboratory

Next Steps Planning is an evolutionary process, which means that a detailed document is often Over the past two decades the scientific obsolete as soon as it is printed. However, method has evolved from a close integration of the broad strokes of a plan should not change experimental and theoretical sciences to include frequently, and this overview of the Laboratory’s the computational and information sciences. In intent for biological signatures research is large measure this evolution has been driven by intended to be general enough that it will complexity of the science enterprise and by the remain serviceable for several years. sheer volume, velocity, and variety of data we now have the tools to acquire. The Science of Signatures is overseen by the Chemistry, Life, and Earth Science Directorate Making sense of this expanded scope in Office. data requires sophisticated integration of Chemistry, Life, and Earth Sciences computational sciences, simulation, and Directorate Office visualization with theory, modeling and Phone: (505) 606-2266 experiment. This powerful approach has been Mail Stop: F629 Testing platforms: engineered human Novel detection: Ag85 is a three protein tissues are being designed to replace complex that is showing promise as an early applied with great success in areas such as Los Alamos National Laboratory, 87545 human and animal tissues. indicator of tuberculosis infection. cosmology, nuclear physics, and climate change. Our intent is to transition lessons learned in The biological signatures component of these other areas to biosurveillance. the Science of Signatures is overseen by Bioscience Division and Theoretical Division in conjunction with the other divisions and program entities listed on page 7.

Bioscience Division Phone: (505) 667-2690 Mail Stop: M888 Los Alamos National Laboratory, 87545

Theoretical Division Phone: (505) 665-4055 Mail Stop: B210 Los Alamos National Laboratory, 87545 Modeling: Ribosome simulations to look for Tools: culturing a complete genome from The full Science of Signatures strategic plan, of new antibiotics. single cells. On a concrete level, our next steps as an which this plan is a subset, is available from any organization will be to refine our internal of these offices. strategy, build partnerships, and invest in the people, equipment, and capabilities required to make our plan a reality. The goals articulated in the previous sections will provide the This document was generated by a subset framework for this process. of the Biological Signatures planning team. Primary authors are Basil Swanson (B-10), The Laboratory welcomes participation Benjamin McMahon (T-6), Nicholas Hengartner from internal groups, teams, and individuals (T-6), Goutam Gupta (B-10), Harshini and we are constantly seeking to build new Mukundan (C-PCS), John Dunbar (B-11), Alina partnerships with external organizations. Deshpande (DSA-3). Should you wish details on the various aspects Document preparation by Josh Smith and of the plan mentioned here, please contact us. Rebecca McDonald, ADCLES communications team.

Uncertainty quantification for flu:an Novel detection: Various new methods of orthogonal array based latin-hypercube pathogen analysis are under development. sample allows more accurate UQ.

18 19 Biological Signatures for National Security Los Alamos National Laboratory

First printing, January 2014.

Los Alamos National Laboratory, an affirmative action/equal Opportunity employer, is operated by Los Alamos National Security, LLC, for the National Nuclear Security Administration of the U.S. Department of Energy under contract DE-AC52-06NA25396.

A U.S. Department of Energy Laboratory LA-UR-14-20499

Sara Del Valle, an epidemiologist with Defense Systems and Analysis Division, stands in front of the PowerWall, which is a stereoscopic theater featuring 24 synchronized projectors and screens that project high resolution, three-dimensional models. The PowerWall is used for a wide variety of modeling and simulation projects.

20