COVID-19 Updates

Visit UC San Diego's Coronavirus portal for the latest information for the campus community.

Seminar Presentations in 2014

December 12, 2014 - Olivier Harismendy, PhD, Assistant Professor, Moores Cancer Center, University of California San Diego
Translation of Cancer Genomic Information
In this lecture, I will present how DNA sequencing can help characterize tumors at the molecular level, identify altered pathways and potential therapeutic targets. I will also review how this information is being used in a clinical setting, highlighting its transformative aspects as well as current limitations.
Bio: Dr. Harismendy graduated with an M.S. in Process Engineering from ENSTA-ParisTech (France) and a joint M.S. in Microbiology from the Pasteur Institute and Paris 7 University. He obtained his PhD in Microbiology from the same university, studying RNA polymerase III transcriptional regulation in yeast using ChIP-chip at Dr Sentenac’s laboratory (CEA-Saclay, France). He then joined Dr. Edelman’s Department of Neurobiology at The Scripps Research Institute (La Jolla, CA), where he developed ChIP-Seq approaches to study Neuron Restrictive Silencing Factor DNA binding in mouse developing brains. Under the mentorship of Dr Kelly Frazer, Dr. Harismendy went on to develop applications of high throughput sequencing for translational research : evaluating methods for targeted sequencing, exploring the role of regulatory variants in common diseases, and detecting and studying the role of somatic mutations in cancer. He joined the UC San Diego Moores Cancer Center in 2009 where he is currently leading the Oncogenomics laboratory. His current research focuses on the development of assays and computational approaches to study tumor heterogeneity predict drug response and understand the contribution of gene regulatory elements in cancer etiology and progression. Strong of his experience in the molecular analysis of clinical samples and the interpretation of the data for cancer care, Dr Harismendy participates in the molecular tumor board and the protocol monitoring and review committee to review and advise on the design and execution of molecularly guided clinical trials.

December 5, 2014 -
Wei Wei, MS, PhD Student, University of California San Diego
Ranking MeSH terms using the factor graph model and the sum product algorithm
Automatically assigning MeSH (Medical Subject Headings) to articles is an active research topic. Recent work demonstrated the feasibility of improving the existing automated Medical Text Indexer (MTI) system, developed at the NLM. Encouraged by this work, we propose a novel data-driven approach that uses semantic distances in the MeSH ontology for automated MeSH assignment. Specifically, we developed a graphical model to propagate belief through a citation network to provide robust MeSH term recommendation. Our preliminary results indicate that this approach can reach high Mean Average Precision in some scenarios.

Eric Levy, BS, PhD Student, University of California, San Diego
Improving predictive models in cancer and Kawasaki disease
Abstract: Of critical interest for clinical applications is the ability to predict disease states for new patients using a combination of clinical and molecular data. In Kawasaki disease, the ability to determine if a new patient is likely to have the disease can prevent delayed diagnoses and earlier treatment, improving disease outcome. In cancer, the ability to stratify patients by predicted outcome, cancer subtypes, or treatment response can greatly aid in the management of the disease for an individual patient. We would like to use a combination of data-driven and model development methods to improve predictive models in these diseases.

November 21, 2014 - Hwanjo Yu, Associate Professor, POSTECH
Search and Mining for Big Data
Big data is recently defined (by Gartner) as high volume, high velocity, and/or high variety information assets that require new forms of processing to enable enhanced decision making, insight discovery and process optimization. In this talk, we first present key challenges in Big data programming, that are distinct from conventional parallel processing. After that, we introduce several research projects dealing with large volume of data in the data mining lab at POSTECH, that are, PubMed relevance feedback search engine, blackbox video search, novel recommendation, and timing when to recommend.
Bio: Prof. Hwanjo Yu received his PhD in Computer Science at the University of Illinois at Urbana-Champaign at June 2004 under the supervision of Prof. Jiawei Han. From July 2004 to January 2008, he had been an assistant professor at the University of Iowa. He is now an associate professor at POSTECH (Pohang University of Science and Technology). He developed influential algorithms and systems in the areas of data mining, database, and machine learning, including (1) algorithms for classifying without negative data (PEBL,SVMC), (2) privacy-preserving SVM algorithms (PP-SVM), (3) SVM-JAVA: an educational java open source for SVM, (4) RefMed: the relevance feedback search engine for PubMed, and (5) TurboGraph: a fast parallel graph engine handling billion-scale graphs in a single PC. His methods and algorithms were published in prestigious journals and conferences including ACM SIGMOD, ACM SIGKDD, IEEE ICDE, IEEE ICDM, ACM CIKM, etc., where he is also serving as a program committee.

November 14, 2014 - Son Doan, PhD, Research Scientist, Division of Biomedical Informatics, University of California, San Diego
Text Mining in Disease Outbreak Surveillance: a Big Data Perspective
Early detection of disease outbreaks is an important feature in any biosurveillance system, helping public health workers quickly response to outbreak for prevention and preparedness. In this talk, we will share experience on building the BioCaster system -an ontology-based text mining system for detecting and tracking infectious diseases from Web data. First, we will present the general architecture as well as the design of the system. The system initially crawls news repots from 1,700 RSS feeds, classifies, semantically analyzes and plots into Google maps. It processes over 9000 news reports a days, consisting of four main stages: topic classification, named entity recognition, disease/location detection and event recognition. The core component of the system is the BioCaster ontology which includes >300 disease names in 12 languages. Second, we will discuss about the Web data ‘quality’ in biosurveillance. We will show that social media such as Twitter, beside official news reports, can be an important resource in biosurveillance systems: it quickly detects disease outbreaks while keeping high accuracy. Several challenges using social media will be discussed. The BioCaster system was being used by CDC, the World Health Organization, the European Centre for Disease Prevention as well as for syndromic surveillance in 2012 Summer Olympics in London.
Bio: Son Doan obtained his PhD degree from Japan Advanced Institute of Science and Technology in 2005 in Computer Science, specializing in natural language processing (NLP) and text mining. He worked as a postdoc from 2006 to 2008 at the National Institute of Informatics (NII) in Tokyo, Japan, focusing on the development of the BioCaster system, a bio-surveillance system for public health. Then he spent two years (2009, 2010) at Vanderbilt University as a Postdoctoral Fellow, working on MedEx – a system for medication extraction from clinical text. He returned to NII in 2010 to work on social data mining using Twitter in public health. Since 2012, he joined the Division of Biomedical Informatics at UC San Diego, working on NLP projects. He is the main developer of the PhenDisco system - a new search system for NCBI’s database of phenotypes and genotypes (dbGaP). His research results using social media in biosurveillance were featured in public media such as Nature, New Scientist, and Asahi Shimbun.

November 7, 2014 - Sándor Szalma, PhD, Head, External Innovation, R&D IT at Janssen Research and Development LLC
The Role of Pre-competitive Translational Informatics in Pharma R&D
Translational medicine and biomarker research is now important part of the pharmaceutical research and development value chain. To support these novel areas, an emerging discipline – translational informatics – has been introduced in the informatics departments or IT groups within pharmaceutical companies. Janssen R&D has also established such a department and in this presentation we will discuss the strategies and approaches this department employing in building up the translational informatics capability and ecosystem with special emphasis on pre-competitive collaborations. Also, open source tools such as tranSMART and NDEx will be highlighted.
Sándor Szalma is head of Translational Informatics and External Innovation, R&D IT in Janssen Research & Development, LLC. He is responsible for implementing the translational informatics capability and establishing and fostering large scale open innovation collaborations to support discovery, translational, clinical and safety science including public-private partnerships and various pre-competitive collaborations such as Innovative Medicines Initiative consortia. His current interests span researching, developing and applying bio- and medical informatics methodologies for enhancing drug discovery and development, biomarkers and translational medicine and utilizing emerging technologies such as social and mobile computing and remote sensors for discovering digital biomarkers, improving clinical development, pharmacovigilance and healthcare. He is member of the board of the Pistoia Alliance, member of the Translational Medicine Advisory Committee of the PhRMA Foundation and co-leading the Data & Knowledge Management Strategic Governance Group of EFPIA. His past positions included president of MeTa Informatics, general manager of QuantumBio and senior director of Computational Biology and Bioinformatics at Accelrys, Inc. He was co-founder of Acheuron Pharmaceuticals, Inc. He lectured at UCSD Extension and now is adjunct professor at Rutgers University in the Computational Biology and Molecular Biophysics program. He is the author of more than 40 scientific publications and book chapters and two patents. He received his doctoral degree in chemistry from A. Szent-Györgyi Medical University in Szeged, Hungary.

October 31, 2014 -
Zachary Lipton
, PhD Student, University of California San Diego
Optimally Thresholding Classifiers to Maximize F1 Measures
F1 measure is widely used to measure the success of a binary classifier when one class is rare. Micro average, macro average, and per instance average F1 measures are used to evaluate multilabel classifiers. We derive the relationship between the best achievable F1 measure and the decision-making threshold that achieves this optimum. If the classifier outputs are well-calibrated conditional probabilities, then the optimal threshold is half the optimal F1 measure. When a classifier is completely uninformative, we prove that the optimal behavior is to classify all examples as positive. If the actual prevalence of positive examples is low, then this behavior is undesirable. As a case study, we discuss the results, which can be far from optimal, of using optimal F1 thresholds when predicting 26,853 labels for biomedical documents from PubMed.

Zhanglong Ji, MS, PhD Student, University of California San Diego
Select and Label (S&L): a Task-Driven Privacy-Preserving Data Synthesization Framework
Abstract: Privacy is a big concern to the public but data sharing has tremendous societal benefits, especially in biomedicine. Existing model perturbation methods can only support a limited number of exploratory model construction before the privacy budget is depleted. On the other hand, most data synthesization approaches are not model specific, which have limited utility for any specific task. We developed a novel differentially private data synthesization framework called select and label (S&L), which can generate synthetic data to meet the classification need. The basic idea is to synthesize ambiguous points near decision boundary (i.e., of the classification model). These points will then be weighted and labeled by a differentially private procedure that is optimized for the classification model. We applied our framework to kernel SVM models and demonstrated superior performance than existing approaches.

October 21, 2014 - Gert Lankriet, PhD, Associate Professor, Jacobs School of Engineering, UCSD
​Music Search and Recommendation from Millions of Songs
Advances in music production, distribution and consumption have made millions of songs available to virtually anyone on the planet, through the Internet. To allow users to retrieve the desired content from this nearly infinite pool of possibilities, algorithms for automatic music indexing and recommendation are a must.
In this talk, I will discuss two aspects of automated music analysis for music search and recommendation: i) automated music tagging for semantic retrieval (e.g., searching for ``funky jazz with male vocals''), and ii) a query-by-example paradigm for content-based music recommendation, wherein a user queries the system by providing one or more songs, and the system responds with a list of relevant or similar song recommendations (e.g., playlist generation for online radio). Finally, I will introduce our most recent research on zero-click recommendation, which leverages various ``sensor'' signals in smartphones to infer user context (activity, mood) and provide music recommendations accordingly, without requiring an active user query (zero click).
I will provide both high-level discussion and technical detail. For example, for query-by-example search, collaborative filter techniques perform well when historical data (e.g., user ratings, user playlists, etc.) is available. However, their reliance on historical data impedes performance on novel or unpopular items. To combat this problem, we rely on content-based similarity, which naturally extends to novel items, but is typically out-performed by collaborative filter methods. I will present a method for optimizing content-based similarity by learning from a sample of collaborative filter data. I will show how this algorithm may be adapted to improve recommendations if a variety of information besides musical content is available as well (e.g., music video clips, web documents and/or art work describing musical artists).
Bio: Gert Lanckriet received a Master's degree in Electrical Engineering from the Katholieke Universiteit Leuven, Leuven, Belgium, in 2000 and the M.S. and Ph.D. degrees in Electrical Engineering and Computer Science from the University of California, Berkeley in 2001 respectively 2005. In 2005, he joined the Department of Electrical and Computer Engineering at the University of California, San Diego, where he heads the Computer Audition Lab and is a co-PI of the Distributed Health Lab. He was awarded the SIAM Optimization Prize in 2008 and is the recipient of a Hellman Fellowship, an IBM Faculty Award, an NSF CAREER Award and an Alfred P. Sloan Foundation Research Fellowship. In 2011, MIT Technology Review named him one of the 35 top young technology innovators in the world (TR35). His lab received a Yahoo! Key Scientific Challenges Award, a Qualcomm Innovation Fellowship and a Google Research Award. In 2014, he received the Best Ten-Year Paper Award at the International Conference on Machine Learning. His research focuses on machine learning, optimization, big data analytics, and crowdsourcing, with applications in music and multimedia search and recommendation.

October 17, 2014 - Kai Wang, PhD, Assistant Professor, Zilkha Neurogenetic Institute, University of Southern California
Bioinformatics approaches for functional interpretation of genome variation
The research in our lab focuses on developing bioinformatics approaches for interpreting the functionality of genetic variants from genome sequencing data. In my talk, I will describe a few ongoing projects: (1) Phenolyzer, which analyzes user-supplied list of phenotype terms and assign most likely candidate genes that are associated with the phenotypes, by integrating multiple sources of gene-pathway-disease-phenotype information. Phenolyzer has been integrated into wANNOVAR, a web server that analyzes genome sequencing data to find disease genes (2) iCAGES (integrated CAncer GEnome Score), which is an effective tool for prioritizing cancer driver genes for a patient using genome sequencing data. iCAGES is implemented with radial Support Vector Machine (SVM) trained on somatic non-synonymous variants from COSMIC and Uniprot databases, followed by a two-step ranking process to employ related biological prior knowledge derived from Phenolyzer. These tools will help researchers better understand the functional consequences of genetic variants identified from genome sequencing studies.​
Bio: Kai Wang is an Assistant Professor in Psychiatry and Preventive Medicine and a Member, Zilkha Neurogenetic Institute and Norris Comprehensive Cancer Center at the Keck School of Medicine of USC. His current research focus is on Develop bioinformatics and computational biology methods to handle high-throughput genomics data sets, especially next-generation sequencing data. Previously, we have worked on SNP arrays and developed the PennCNV software for identifying copy number variations (CNVs) and the GenGen software for pathway-based association tests. Recently, we focused on whole-genome and whole-exome sequencing data, and developed the ANNOVAR software for functional annotation of genetic variants. He also applies genomic approaches to a variety of brain disorders, especially brain cancer and neurodevelopmental diseases. He received his B.S. from Peking University in 2000 and Ph.D. from University of Washington in 2005 and worked as a post-doc at the University of Pennsylvania and Children's Hospital of Philadelphia until 2010.

October 10, 2014 - Naveen Ashish, PhD, Associate Professor of Research, Laboratory of Neuro Imaging, University of Southern California
Data Integration of Alzheimer’s Disease Data: Experiences in the GAAIN Project
In this talk I will present our work in data integration in the context of ongoing “GAAIN’ project. GAAIN – The Global Alzheimer’s Association Interactive Network is building an integrated network of Alzheimer’s research data where researchers can get seamless access to harmonized Alzheimer’s data contribute by different research groups across the globe. I will begin with an overview of GAAIN, and the talk is then focused on two specific research areas within. The first is our current on data element mapping of Alzheimer’s datasets. We will describe our work on the GAAIN-Mapper, which is an intelligent software assistant to assist developers with the task of identifying and matching data elements in a new (Alzheimer’s) dataset to elements in the GAAIN common data model. Our approach significantly leverages the available dataset documentation, typically in the form of Data Dictionaries associated with the data. I will also describe our work on the GAAIN Automated Data Transformer, which is a declarative data transformation system to transform a new dataset into the GAAIN common model representation.
Bio: Naveen Ashish is an Associate Professor of Research Neurology (Informatics) at the Keck School of Medicine of USC. His current research focus is on medical data integration, particularly the GAAIN project on Alzheimer’s disease data integration. He has published extensively in the data integration area and has also authored two books on the topic. He received his PhD in Computer Science from the University of Southern California in 2000 and a BTech, also in Computer Science, from IIT Kanpur (India) in 1993.

​October 3, 2014 - Richard S Garfein, PhD, MPH, Professor of Global Public Health, University of California San Diego.
Video-Directly Observed therapy (VDOT): a solution for monitoring TB treatment adherence
Background: Over 8.8 million people become ill and 1.4 million people die annually from tuberculosis (TB) worldwide. TB is treatable with antibiotics; however, poor adherence to daily medication regimens lasting >6 months leads to ongoing disease transmission, higher mortality, and development of antibiotic resistant bacteria. “Directly observed therapy" (DOT) is recommended to minimize these problems by having health workers watch patients take each medication dose. While effective, DOT is costly, time consuming, invasive, and impractical for some patients. Thus, some TB programs have begun teleconferencing with patients using analogue or digital devices. However, DOT records are still maintained in paper form or local databases making analysis of medication adherence and outcomes difficult.
Intervention or response: We developed and pilot-tested the Video DOT (VDOT) System whereby patients use mobile phones to record and securely transfer time-stamped videos of them taking their medications, which are then viewed remotely by a health worker who documents doses observed in the system's database. TB patients were enrolled in San Diego, California (43) and Tijuana, Mexico (9) to test the feasibility and acceptability of VDOT. Participants were interviewed at baseline and at treatment completion.
Results and lessons learned: Overall, participants ages ranged from 18 to 86 years, 54% were male, and 77% were non-white. Participants used VDOT a mean of 5.5 months (range: 1-11 months); 7 (13%) returned to in-person DOT. Treatment adherence was excellent in San Diego (93%) and Tijuana (96%). Post-treatment interview responses were similar across cities. Overall, 89% of patients reported never/rarely having problems recording videos, 92% preferred VDOT over in-person DOT, 81% thought VDOT was more confidential, and 100% would recommend VDOT to others. Three non-compliant participants were returned to in-person DOT, suggesting the need for both options. Data collected in this common database were easily analyzed.
Conclusions: VDOT provides a promising mobile solution to the high cost and burden of in-person DOT for monitoring TB and other conditions that require strict treatment adherence. Once scaled, standardized medication adherence data from multiple TB control programs may be analyzed quickly and efficiently. Data may also be used to generate health insurance reimbursement requests as telehealth becomes more accepted by insurers.
​Bio: ​​Dr. Garfein is a Professor in the Division of Global Public Health, Department of Medicine at UCSD. He trained as an infectious disease epidemiologist, first earning an M.P.H. from the San Diego State University, School of Public Health in 1989 and then a Ph.D. from Johns Hopkins University, School of Hygiene and Public Health in 1997. Between his graduate training and joining the UCSD faculty in 2005, Dr. Garfein served as an Epidemic Intelligence Service Officer in the Division of Viral Hepatitis and a Senior Staff Epidemiologist in the Division of HIV/AIDS Prevention at the Centers for Disease Control and Prevention. His research interests involve identifying risk factors for and developing interventions to prevent infectious diseases associated with substance abuse.

 June 6, 2014 – Dexter Pratt - Director, NDEx project, University of California, San Diego. NDEx, the Network Data Exchange: Novel Infrastructure for Collaboration, Crowdsourcing, and Application Development with Biological Networks 

Abstract: The NDEx Project is creating a "Network Commons", an open-source software system to facilitate collaboration, publication, and application development for scientists and organizations working with biological networks of multiple types and in multiple formats. An NDEx-based public portal will create a "Google Docs" for scientists with networks. This talk presents key features of the NDEx infrastructure and the forthcoming public portal, and will also discuss the facilities of the existing beta system, the integration of Cytoscape and the NDEx REST server, and the treatment of provenance information in the NDEx data model. The status of NDEx in collaborations with other organizations and the use (or development) of standards will be summarized. 
Bio: Dexter Pratt is the Director of the NDEx Project, a project of the Ideker Lab at UCSD and the Cytoscape Consortium ( He is the creator of BEL (the Biological Expression Language) and has the role of Founder and Evangelist in the OpenBEL Consortium ( In recent years he served as VP Knowledge and Innovation at Selventa Inc., where he led teams in algorithm development, curation, and drug discovery consulting engagements. His earlier career was in commercial artificial intelligence research and product development at companies including Lisp Machine Inc. and the CYC Project at the Microelectronics and Computer Consortium. Dexter will also be leading a UCSD team in a DARPA-funded collaboration with the Sorger lab at HMS, linking large, qualitative biological models with executable quantitative models. BEL has been selected for use and extension in this project because of its focus on the representation of scientific findings, biological context, and the encoding of qualitative causal relationships.

May 30, 2014 – Enming Luo - Ph.D. Student, Department of Electrical and Computer Engineering, University of California, San Diego.
Truong Nguyen, Ph.D. - Professor, Department of Electrical and Computer Engineering, University of California, San Diego.
Image Denoising by Targeted External Databases
Abstract: Classical image denoising algorithms based on single noisy images and generic image databases will soon reach their performance limits. We envision that future image denoising should be target-oriented, i.e., for specific objects to be denoised, only similar images should be used for training. We thus propose to denoise images using targeted external image databases. Formulating denoising as an optimal filter design problem, we utilize the targeted databases to (1) determine the basis functions of the optimal filter by means of group sparsity; (2) determine the spectral coefficients of the optimal filter by means of localized priors. For a variety of scenarios such as text images, multiview images, face images and medical images, we demonstrate superior denoising results over existing algorithms.
Bio: Enming Luo is a fourth year PhD student, supervised by Prof. Truong Nguyen in the ECE department at UCSD. His research is on ill-posed inverse problems in signal processing. Some applications include image denoising, super-resolution and deblurring. Prior to his study at UCSD, he got his master degree from HKUST, where he worked on video compression. Truong Nguyen manages the Video Processing Group at UCSD. The group's research objective is to generate enabling technology for low-cost mobile devices. The research involves the invention, development, analysis and implementation of multirate systems with emphasis on low-power application in image and video processing. 

May 23, 2014 – Ko-Wei Lin, DVM, PhD - Postdoctoral Fellow, Division of Biomedical Informatics, University of California San Diego
Let's groove with PhenDisco: a new phenotype search tool for the database of Genotypes and Phenotypes (dbGaP)

Abstract: The database of Genotypes and Phenotypes (dbGaP), developed by the National Center for Biotechnology Information (NCBI), contains data generated from genomics studies such as genome-wide association studies (GWAS). These data can be re-used to facilitate novel scientific discoveries and cross-study validations, and to reduce cost and time for exploratory research. However, idiosyncrasies and inconsistencies in phenotype variables are a major barrier for researchers to reuse these data. In PFINDR (Phenotype Finder IN Database Resources) project, we developed PhenDisco (Phenotype Discoverer,, a new phenotype information retrieval system for dbGaP. This presentation will discuss how we standardize phenotype variables in developing PhenDisco tool. 
Bio: Dr. Lin is a Postdoctoral Fellow at Division of Biomedical Informatics, UCSD. Dr. Lin received DVM degree from National Taiwan University, and Ph.D. in Comparative Biomedical Science focusing on Cell Biology with a minor in Biotechnology from North Carolina State University. Her research was discovering pathogenic mechanisms of lung inflammation in Chronic Obstructive Pulmonary Disease (COPD) and asthma, specifically focused on signaling pathways and proteomics. Prior to joining DBMI, she was a Postdoctoral Scholar in the Division of Pulmonary and Critical Care Medicine at UCSD, and worked on examine immune pathways in asthma and acute lung injury, specifically focused on T cells, miRNAs, and interactions between innate and adaptive immunity. Dr. Lin joined Phenotype Finding IN Data Resources (PFINDR) team at DBMI from 2011. Her current research focuses on phenotype standardization in developing information retrieval tools for genomics database. 

May 16, 2014 – Kai Zheng, PhD - Associate Professor, School of Public Health Department of Health Management and Policy, Associate Professor, School of Information, University of Michigan 
Computational Ethnography: Automated and Unobtrusive Means for Collecting Data in situ for Human–Computer Interaction Studies 
Abstract: Human–computer interaction (HCI) studies in healthcare have been traditionally conducted in the forms of expert inspection, usability experiments, field studies, and direct perception solicitation through questionnaire surveys, interviews, or focus groups. While these traditional HCI approaches are of great merit and are indispensable in studying and improving usability of software systems and medical devices in healthcare, they have several major limitations in common. In this talk, I will introduce an emerging family of methods for conducting HCI studies in healthcare, computational ethnography, which leverages automated and unobtrusive or less obtrusive means for collecting in situ data reflective of real end users’ actual, unaltered behavior of using a software system or a device in real-world settings. These methods are based on the premise that user interactions with modern technologies, or interpersonal communications they conduct mediated by modern technologies, always leave some “digital traces” behind that can be utilized by HCI experts to fully or partially re-enact the activities. In the talk, I will introduce the definition of computational ethnography, common types of digital trace data that are either being routinely collected in a healthcare environment or can be proactively collected by HCI researchers, and commonly used analytical approaches for making sense of such data. I will conclude the talk with two use cases illustrating how this new family of methods has been applied in healthcare to study end users’ interactions with technological interventions in their everyday routines. 

Bio: Kai Zheng is jointly appointed as Associate Professor of Health Management and Policy in the School of Public Health and Associate Professor of Information in the School of Information at the University of Michigan. He is also affiliated with the University of Michigan’s School of Nursing, Michigan Institute for Clinical and Health Research, Medical School Department of Computational Medicine and Bioinformatics, and Center for Entrepreneurship. He co-directs the Bio-Repository and Biomedical Informatics Core of the University of Michigan Health System and Peking University Health Science Center Joint Institute for Translational and Clinical Research. Zheng received his PhD degree from Carnegie Mellon University where his dissertation won the University’s 2007 William W. Cooper Doctoral Dissertation Award in Management or Management Science. He is the recipient of the 2011 American Medical Informatics Association New Investigator Award that recognizes early informatics contributions and significant scholarly achievements on the basis of scientific merit and research excellence. 

May 9, 2014 – Peter W. Rose, PhD - Scientific Lead, RCSB Protein Data Bank, San Diego Supercomputer Center, University of California, San Diego. 
Biology in 3D -- The RCSB Protein Data Bank 
Abstract: The Protein Data Bank (PDB) is the single worldwide repository of experimentally determined structures of proteins, nucleic acids, and complex assemblies. Dr. Rose will give an overview of the PDB, recent trends, and discuss its application to drug discovery. 
Bio: Dr. Rose is the Scientific Lead of the RCSB Protein Data Bank at the University of California, San Diego. He is responsible for the scientific objectives and further development of the RCSB PDB. Prior to joining UCSD, he held research and management positions of increasing responsibility at Pfizer Global R&D La Jolla., formerly Agouron Pharmaceuticals. Peter was instrumental in the establishment of the structure-based design platform at Agouron and its global adoption by Pfizer. As Director of Computational Chemistry and Bioinformatics, he oversaw Structural Bioinformatics, Structure-Based Drug Design, and Scientific Computing groups. He was a postdoctoral researcher at the University of California Irvine and received his Ph.D. in 1990 from the Technical University of Munich. 

May 2, 2014 – Tzyy-Ping Jung, PhD - Co-Director, Center for Advanced Neurological Engineering; Associate Director, Swartz Center for Computational Neuroscience; Professor, Dept. of Bioengineering, University of California, San Diego. 
Ubiquitous Monitoring of Brain/Body Dynamics via Multi-Tiered Cloud Computing 
Abstract: Late twentieth and early twenty first centuries have witnessed a revolution in neuroscience through instrumentation, signal processing, and computer and information technologies. This revolution, multi-disciplinary collaboration and innovations have made big impacts on medical research. A natural next step is to create the ability to leverage the vast worldwide neuroscience efforts to further advance neuroscience research, and improve prevention, diagnosis, and treatment of neurological diseases and injuries. NCTU and UCSD have jointly developed miniature sensors and sensing systems that allow noninvasive, nonintrusive and continuous monitoring of the brain and body activities of unconstrained, freely moving participants in real-work environments such as the home, shopping malls, movie theaters, etc. We have also made a lot of progresses on developing mathematical modeling methods and software to find statistical relationships between moment-to-moment variations in environmental, behavioral, and functional brain dynamic recordings. Furthermore, we have jointly developed a pervasive brain monitoring and data sharing infrastructure featuring wireless headgears, MEMS motion sensors, smart phones and cloud servers. With the aid of machine-to-machine (M2M) communication, multi-tier fog/cloud computing and linked data technologies, we succeeded in disseminating real-time mobile brain and body data streams simultaneously to multiple sites around the globe using state-of-art publish/subscribe mechanisms. We are also working on meta-data models for our devices and data archives so that researchers can search for and connect to these on-line resources through the Linked Data Cloud. My presentation will provide an overview of system architecture and its capability. 
Bio: Prof. Jung received the B.S. degree in electronics engineering from National Chiao Tung University, Taiwan, in 1984, and the M.S. and Ph.D. degrees in electrical engineering from The Ohio State University in 1989 and 1993, respectively. During 1993-1996, he was a Research Associate of the National Research Council of the National Academy of Sciences working at the Computational Neurobiology Laboratory, The Salk Institute for Biological Studies, La Jolla, CA. He is currently a Research Scientist at the Institute for Neural Computation of University of California (UCSD), San Diego. He is also a Co-Director of Center for Advanced Neurological Engineering and Associate Director of the Swartz Center for Computational Neuroscience and an Adjunct Professor of Department of Bioengineering at UCSD. In addition, he is an Adjunct Professor of Department of Computer Science at National Chiao-Tung University. Dr. Jung’s research emphasis has been placed on integration of the cognitive sciences, basic sciences, and engineering in service of insights into functions of brain, cognition, and behavior. In particular, he and colleagues have proposed the applications of blind source separation techniques to multichannel EEG/MEG/ERP and fMRI data to separate different brain processes into statistically independent components arising from functionally distinct brain or extra-brain networks. This signal decomposition would not have obtained from any conventional signal-processing approaches. The new signal-processing technique opens a novel and revolutionary window into complex event-related brain data (EEG, MEG, ERP, fMRI, etc.) that leads to a more detailed understanding of the strengths and limitations of the human mind, plus possible applications to medicine and to cognitive testing and monitoring. 

April 25, 2014 – Richard Schwab, M.D., - Associate Clinical Professor, Division of Hematology-Oncology and Regulatory Director, UCSD Moores Cancer Center Biorepository, University of California, San Diego. 
Deciphering Breast Cancer with DNA Sequencing 
Abstract: Breast cancer is a heterogeneous disease and the solid tumor with the most useful predictive cancer biomarkers; Estrogen Receptor and HER2. Dr. Schwab with discuss recent work at UCSD leveraging advances in next-generation sequencing to identify predictive biomarkers and to further our understanding of breast cancer. 
Bio: Dr. Schwab earned his M.D. degree from Albert Einstein College of Medicine in 2000. He served his residency and fellowship in Internal Medicine and Hematology/Oncology in 2003 and 2006, respectively, both at UCSD. Dr. Schwab’s research focuses on biomarkers to diagnosis and refine the treatment of cancer. His current projects include precision oncology using DNA sequencing, target discovery using RNA sequencing, and validation of novel cancer initiating cell markers. In supporting these and future biomarker studies Dr. Schwab co-directs a biorepository to prospectively collect de-identified blood, urine and tumor specimens. 

April 18, 2014 - Jason C. Liang, Ph.D. - Founder and Principal at Liang-Zhi International Intellectual Property Company. Hsun-Hsien “Shane” Chang, MBA, Ph.D. – Scientific Advisor of Wilson Sonsini Goodrich & Rosati Professional Corporation; Adjunct faculty of Harvard Medical School; and Founder of HARVMIT Corp
From Science to Business — How to Devise a Global IP Strategy
Abstract: The presentation is directed to general discussion on how to devise a global IP strategy from the angle of scientists who will take an important part in a biotech company. Specifically, the topic is directed to how your idea gets to the company, objectives of an IP program in a startup company, forms of IP protection, how to devise global IP strategies based on costs and company needs and the overall considerations. 
Bio: Dr. Jason C. Liang, Jason Liang is the founder of Liang-Zhi International IP firm and a member of Acuity Law Group. Before Liang-Zhi, Jason worked 4 years at WSGR, a law firm that is known to serve startup companies for their legal needs. Before joined WSGR, Jason was a patent agent at Foley and Lardner PC, an international full service law firm that is known to serve big pharma and biotech companies. Dr. Liang has strong connection with Asian markets and has provided his patent services to several biotech companies in Taiwan and China. Jason also worked as an in house IP manager at Optimer Pharmaceuticals until her IPO and he holds a Ph.D. in chemistry and acquired J.D. last year with the emphasis on intellectual property laws. Dr. Hsun-Hsien Shane Chang graduated from Carnegie Mellon University with PhD in Electrical and Computer Engineering in 2007 and went on to serve as a Research Associate in the Division of Health Sciences and Technology, Harvard Medical School and MIT until 2013. He also earned a MBA degree from Taiwan in 2000. He is now a Scientific Advisor of WSGR and still holds adjunct faculty position at Harvard Medical School. He founds HARVMIT Corp. recently in San Diego. Dr. Chang 

April 11, 2014, Steven Stupp, PhD - CEO, Trigeminal Solutions Inc. 
Assumption-Free Feature Extraction in Noisy and Extremely Underdetermined Medical Datasets
Abstract: Leveraging data to make useful predictions is not new, but it has taken on greater prominence with the rise of big data. This challenge is compounded in so-called 'underdetermined' problems in which there are a very large number of possible predictors (or features) and a limited (and much smaller) number of observations, which are very common in medicine and genetics. Because these problems do not have unique solutions, how can you avoid the curse of dimensionality, address multiple testing/false positives, and avoid overfitting to successfully identify predictors? Said differently, how can you find the right need needle in a huge stack of needles? Historically, feature-selection and extraction techniques attempt to simplify underdetermined problems by discarding (hopefully) irrelevant features, thereby reducing the size of the search space. Trigeminal Solutions, a healthcare-technology company based in the San Francisco Bay area, has developed a novel feature-extraction technique (Feature Extraction Leveraging Interactions and NoisE or FELINE) that takes the seemingly paradoxical approach of expanding the size of the search space in underdetermined problems. FELINE has been successfully applied to very underdetermined problems in medicine and genetics. For example, in a clinical trial we used FELINE to successfully identify individual-specific migraine triggers, which help predispose subjects to migraine headaches. We have also successfully applied FELINE to the Genome Association Information Network genetic dataset for major depressive disorder (MDD). In each of these cases, FELINE identified known and previously unreported features, which were subsequently confirmed. For MDD, this involved considering more than 1.4 trillion possible predictors for 3,360 subjects, a ratio of more than 400 million to one. FELINE is an empirical technique with no assumptions and no arbitrary adjustable parameters. As such, it is potentially a powerful and general technique that can be applied to a wide variety of problems, including chronic diseases with episodic manifestations, genetic datasets for other diseases, and electronic medical records. 
Bio: Steven Stupp has more than 20 years experience in the technology industry, including founding and serving on the management teams of several startups (as well as Trigeminal Solutions), and contributing in technical fields such as superconductivity, digital communications, and biosensors, at organizations such as Philips Research, Quantum Corp. and Stanford University. Steve has M.S. and Ph.D. degrees in condensed matter physics from the University of Illinois, and a B.S. degree in physics, summa cum laude, from Tufts University. He is a registered patent agent who has written more than 500 patents for startups and Fortune 500 companies, including the first patents for Apple, Inc.'s iPhone. 

April 4, 2014, James Killeen MD - Associate Clinical Professor of Emergency Medicine, Department of Emergency Medicine; Department of Hyperbaric Medicine, UC San Diego. Edward M. Castillo, M.P.H., Ph.D., Assistant Adjunct Professor, Department of Emergency Medicine, UC San Diego. 
Integrating Environmental Data into a Personal Health Record for Asthma Patients 
Abstract: the purpose of this project was to test and demonstrate secure and reliable bidirectional transport of patient data between end-users of National Association for Trusted Exchange (NATE) community Health Information Service Providers (HISPSs) and patient-owned personal health records (PHRs). The project leveraged existing community resources to test data that are in a standard structured exchange format from an existing community Health Information Exchange (San Diego Beacon Community) as well as new data from innovative environmental & patient monitoring systems (DELPHI Project and CitiSense) to assess scalability and interoperability. For this project we utilized use cases focused on the transport of data for patients with asthma. 
Bio: Dr. James Killeen, MD is a UCSD Associate Professor for both the Emergency Department and Hyperbaric Medicine Department. He is also Director of Information Technology Services for the UCSD Emergency Department. Dr. Killeen is the designer and developer of the electronic health record (EHR) called “WebCHARTS.” This copy written software package is designed for the emergency department featuring patient documentation, physician order entry and patient tracking. He is working with CHIP and the SD county on creating a county wide patient identifier and referral system for patients with no medical home called “Safety Net Connect” based off a prior pilot study called IMPACT ED funded by the Alliance Healthcare Foundation. He is also a member of the software design team for the National Library of Medicine funded “WIISARD” and “WIISARD SAGE” projects designed for Pre-hospital mass casualty disaster events that include wireless patient tracking system and prehospital provider EHR for on scene patient care. Dr. Killeen is an editor for both the Journal of Emergency Medicine as well as the Western Journal of Emergency Medicine. Dr. Edward Castillo graduated with and MPH from the Graduate School of Public Health in 1999 and went on to earn a PhD in Public Health (Epidemiology) from the San Diego State University/University of California, San Diego Joint Doctoral Program in Public Health in 2003. Prior to joining the Emergency Medicine faculty in 2006, he served as an Epidemiologist with the County of San Diego Health and Human Services Agency, Division of Emergency Medical Services from 2001–2004 and a Research Scientist with the Institute for Public Health at San Diego State University from 2004–2006. His research interests include healthcare utilization, informatics and dietary supplement use. 

March 14, 2014, Bo Dagnall, M.S. - Chief Technologist of VA Account, Enterprise Services, Hewlett Packard 
Applied Health Informatics 
Abstract: The objective of this presentation is to demonstrate how health informatics is applied in industry. I will discuss my journey from a UCSD undergraduate to a 15+ year career in informatics. We will review industry definitions, business problems and the core components of informatics. As informatics is at the center of a major investment in health IT largely funded by the US government, I will highlight some of the central policies, legislation, industry bodies, and product companies influencing the future. We will also showcase the Veteran’s Health Administration as a case study for applied health informatics discussing their journey and findings over the last 20+ years. 
Bio: I am the Chief Technologist within HP Enterprise Services for our Department of Veteran’s Affairs account. The Veteran’s Health Administration has been a recognized leader in health informatics and health IT for over 20 years. Our VHA account has over 700 employees working on numerous projects including major enhancements to VistA – the electronic health record system created by the VA in the 1990’s. I oversee technology strategy, innovation, effective use of technology for these projects. 
I am formally trained in health informatics with a MS in Medical Informatics from the University of Utah and undergraduate degree from UCSD. Throughout my career, I have tended to focus on the technology and information architecture aspect of informatics working originally as a developer and architect of health IT solutions. Career achievements include creating the Health Data Repository for the VHA and defining the eHealth architecture for Queensland Health in Australia. I am also on the Architecture Review Board for HL7. 

March 7, 2014, Wei Wang, PhD - Professor, Department of Computer Science, University of California, Los Angeles 
Detecting and Correcting Spurious Transcriptome Inference due to RNAseq Reads Misalignment 
Abstract: RNA-seq techniques provide an unparalleled means for exploring a transcriptome with deep coverage and base pair level resolution. Various analysis tools have been developed to align and assemble RNA-seq data, such as the widely used TopHat/Cufflinks pipeline. A common observation is that a sizable fraction of the fragments/reads align to multiple locations of the genome. These multiple alignments pose substantial challenges to existing RNA-seq analysis tools. Inappropriate treatment may result in reporting spurious expressed genes (false positives), and missing the real expressed genes (false negatives). Such errors impact the subsequent analysis, such as differential expression analysis. In our study, we observed that about 3.5% of transcripts reported by TopHat/Cufflinks pipeline correspond to annotated nonfunctional pseudogenes. Moreover, about 10.0% of reported trascripts are not annotated in the Ensembl database. These genes could be either novel expressed genes or false discoveries. We examined the underlying genomic features that lead to multiple alignments and investigate how they generate systematic errors in RNA-seq analysis. We developed a general tool, GeneScissors, which exploits machine learning techniques guided by biological knowledge to detect and correct spurious transcriptome inference by existing RNA-seq analysis methods. GeneScissors can predict spurious transcriptome calls due to misalignment with an accuracy close to 90%, which represents substantial improvement over the widely used TopHat/Cufflinks or MapSplice/Cufflinks pipelines.
Bio: Wei Wang is the director of the Scalable Analytics Institute (ScAI). She received a MS degree from the State University of New York at Binghamton in 1995 and a PhD degree in Computer Science from the University of California at Los Angeles in 1999. She was a professor in Computer Science and a member of the Carolina Center for Genomic Sciences and Lineberger Comprehensive Cancer Center at the University of North Carolina at Chapel Hill from 2002 to 2012, and was a research staff member at the IBM T. J. Watson Research Center between 1999 and 2002. Dr. Wang's research interests include big data, data mining, bioinformatics and computational biology, and databases. She has filed seven patents, and has published one monograph and more than one hundred research papers in international journals and major peer-reviewed conference proceedings. 

February 28, 2014, Jason N. Doctor, PhD - Associate Professor, Clinical Pharmacy and Pharmaceutical Economics and Policy,University of Southern California 
Applications of Nudges at the Point-of-Care 
Abstract: The standard model of physician decision-making holds that physicians will generally conform to the technical standards of the profession through the receipt of specialized knowledge that occurs over long and intensive academic preparation. Yet often physicians do not meet these standards and this can have severe public health consequences. For example, in the U.S. half of the 41.2 million adult acute respiratory infection prescriptions are unnecessary (i.e., applied to diagnoses for which there is no evidence of benefit). Negative outcomes of this practice include adverse drug reactions, increased health care costs, and bacterial antibiotic-resistance. Despite educational efforts to curb inappropriate prescribing, physicians continue to prescribe antibiotics at higher rates than necessary. The behavioral economic model of physician decision-making is a descriptive theory that holds that physician decisions making may be affected by a broad set of motivations that in addition to training-based motives may include: Pride in performance, a commitment to consistency, concern over the opinion or approval of peers or senior staff. Further, decisions may be affected by psychological biases and heuristic (rule of thumb) approaches to reasoning. I will discuss our recent efforts to change physician behavior and identify the consistency of these efforts with behavioral economic principles as well as their effectiveness. Bio: Dr. Doctor received his PhD in clinical psychology from the University of California at San Diego. He served as a research fellow at the University of Washington. Prior to coming to USC this year, Dr. Doctor was an Associate Professor of Medical Education & Biomedical Informatics, Health Services and Rehabilitation Medicine at the University of Washington. 

February 21, 2014, William Hsu, PhD - Assistant Professor, Medical Imaging Informatics Group, Department of Radiological Sciences, University of California, Los Angeles 
The Role of Imaging Informatics in Deep Phenotyping 
Abstract: Imaging plays a critical role in illuminating in vivo both the human condition and disease. New advances in imaging technologies have led to major insights into a range of medical conditions, revealing matters of structure and function. Imaging informatics research encompasses an entire spectrum from low-level concepts (e.g., image feature extraction) to higher-level abstractions (e.g., semantic labeling of image regions). In this talk, I will motivate the need for imaging-based observational databases that associate images with other clinical data sources to contextualize observations and reach conclusions about a disease and its evolution. I will describe our efforts to integrate imaging with clinical and –omics level data and discuss challenges related to data quality and predictive modeling.
Bio: William Hsu is an Assistant Professor with the UCLA Medical Imaging Informatics Group in the Department of Radiological Sciences. He received his PhD in Biomedical Engineering with an emphasis in Medical Imaging Informatics from the University of California, Los Angeles in 2009 and a BS degree in Biomedical Engineering from Johns Hopkins University in 2004. His research interests include medical data visualization, disease modeling, knowledge representation, and imaging informatics. He was recently awarded a grant from the American College of Radiology and serves on the advisory board of the American Medical Informatics Association's Biomedical Imaging Informatics Working Group. 

February 14, 2014, Yu-Tsueng Liu, MD, PhD - Assistant Professor of Medicine, Division of Infectious Diseases, Moores Cancer Center, University of California, San Diego 
Standardization of 3D clinical photography: Experience in Mozambique 
Abstract: Kaposi’s sarcoma (KS) is the most frequently occurring cancer in Mozambique among men and the second most frequently occurring cancer among women. Effective therapeutic treatments for KS are poorly understood, as the KS response may be different between areas such as Southeast Africa and western countries where it has been more rigorously studied. Infrastructural inadequacies in the provisions of cancer care in the African region pose great challenges for precise therapeutic monitoring. Therefore, there is an unmet need to develop a simple but accurate tool for improved monitoring and diagnosis in a resource-limited setting. In response, we have created a portable, user-friendly, and low cost 3D imaging system to capture standardized high-quality images of KS lesions using a commercially available camera, a custom adaptor and computational programs. We are currently collaborating with clinicians in Mozambique to test the feasibility of this device for clinical applications.

February 7, 2014, Lucila Ohno-Machado, MD, PhD, MBA - Professor and Division Chief, Division of Biomedical Informatics, Department of Medicine, University of California, San Diego
Introduction to pSCANNER: Patient-Oriented SCAlable National Network for Effective Research 
Abstract: pSCANNER is part of the recently formed PCORnet, a PCORI national network composed of learning healthcare systems and patient-powered research networks. It is designed to be a stakeholder-governed federated network utilizes a distributed architecture to integrate data from three existing networks covering over 21 million patients: (1) the University of California Research eXchange (UC-ReX) network, with data from UC Davis, Irvine, Los Angeles, San Francisco, and San Diego; (2) VA Informatics and Computing Infrastructure (VINCI), with data from Veteran Health Administration’s (VHA) 151 inpatient and 827 outpatient facilities; and (3) the SCAlable National Network for Effectiveness Research (SCANNER), a consortium of three ambulatory care systems in the Greater Los Angeles area supplemented with claims and health information exchange (HIE) data, led by the University of Southern California. We will focus on three conditions: (1) Congestive heart failure, (2) Kawasaki Disease, and (3) Obesity. Patients, patient advocates, domain experts for these conditions, health services researchers, clinicians, and administrators from across the country will participate in pSCANNER’s governance. In addition, we will use innovative user-friendly, online software to conduct a rigorously designed Delphi consensus process that engages patients, clinicians, and researchers in the prioritization of research questions. We will use a privacy-preserving distributed computation model with synchronous mode and asynchronous modes. The distributed system is based on a common data model and allows the construction and evaluation of distributed multivariate models for a variety of statistical analyses. 
Bio: Lucila Ohno-Machado, MD, MBA, PhD received her medical degree from the University of Sao Paulo and her doctoral degree in medical information sciences and computer science from Stanford. She is Associate Dean for Informatics and Technology, and the founding chief of the Division of Biomedical Informatics at UCSD, where she leads a group of faculty with diverse backgrounds in medicine, nursing, informatics, and computer science. Prior to her current position, she was faculty at Brigham and Women’s Hospital, Harvard Medical School and at the MIT Division of Health Sciences and Technology. Dr. Ohno-Machado is an elected fellow of the American College of Medical Informatics, the American Institute for Medical and Biological Engineering, and the American Society for Clinical Investigation. She serves as editor-in-chief for the Journal of the American Medical Informatics Association since 2011. She directs the NIH-funded iDASH National Center for Biomedical Computing. 

January 31, 2014, Charles Elkan, PhD - Professor, Department of Computer Science and Engineering, University of California, San Diego 
The Past and Future of Data Science, Inside and Outside Medicine 
Abstract: This talk will discuss some of the foundations of the field of data science, and in particular will describe a prescient application from 1948. The talk will link this pioneering application to new methods in reinforcement learning, that will, we can hope, enable disruptive advances in clinical medicine and elsewhere. The talk will also discuss the history and future of sentiment analysis, in the context of new ideas for text mining and search engine design. 
Bio: Dr. Charles Elkan is a professor in the Department of Computer Science and Engineering at the University of California, San Diego. In the past, he has been a visiting associate professor at Harvard and a researcher at MIT. Dr. Elkan is known for his academic work in machine learning and data mining. The MEME algorithm developed by his group has been used in over 2000 research projects in biology and computer science. He has won several best paper awards and data mining contests. His M.S. and PhD students have become leaders at many companies and universities. 

January 24, 2014, Xianghong Jasmine Zhou, PhD - Associate Professor, Computational Biology and Bioinformatics Program, Department of Biological Sciences, University of Southern California
Make Big Data Useful: Horizontal and Vertical Data Integration to Study Genes, Networks and Diseases 
Abstract: The life sciences are becoming a big data enterprise with its own data characteristics. To make big data useful, we need to find ways of dealing with the heterogeneity, diversity, and complexity of the data, to identify problems that cannot been solved before, and to develop methods to solve those new problems. In this talk, I will outline a set of novel biological problems that we proposed and solved by integrating a large amount of genomic data in public repositories. We define the integration of the same type of data (e.g. gene expression data) to be horizontal data integration, and that of different type of data (e.g. epigenetic data, gene expression, and genome structures) to be vertical data integration. For the horizontal data integration, I will discuss our recent work on integrating many RNA-seq datasets to perform high-resolution functional annotation of human genome, namely, predicting the functions of individual transcript isoforms. I will also briefly describe our effort in transform public gene expression repositories into a disease diagnosis database. For the vertical data integration, I will discuss our analysis of TCGA data, where we defined the term “multi-dimensional modules” to describe the multi-layer coordinated perturbation on cancer pathways, and describe novel methods to identify such cellular activities. Finally, I will report our most recent work on integrating the 3D chromatin structures, epigenetic modification, and transcription factors to study gene regulation. 

January 17, 2014, Sarah S. Murray, PhD - Associate Professor, Department of Pathology, Director, Genomic Technologies, Center for Advanced Laboratory Medicine, University of California, San Diego 
Clinical Genomics 
Abstract: Advances in sequencing technologies have enabled cost-effective sequencing of large numbers of targeted genes or genomic regions, exomes and even whole genomes both quickly and accurately. The generation of massively parallel sequence data requires sophisticated informatics infrastructure as well as bioinformatics expertise to manage the large amount of data generated as well as process, annotate and interpret complex sequence data into a clinically actionable result. The UCSD clinical genomics lab, part of UCSD’s Center of Advanced Laboratory Medicine, has implemented the use these technologies in a clinical test. The first test, a tumor profiling panel, interrogates key exons of 47 clinically actionable genes in cancer. We have implemented an integrated process encompassing order entry, specimen acquisition, laboratory and bioinformatics analyses, data interpretation, and result reporting, with data flowing between the hospital information system (HIS) and laboratory information system (LIS). The clinical genomics lab is continuing to validate new tests using a similar strategy in order to provide enhanced diagnostic, therapeutic, and prognostic assessment through advanced genomic profiling of a wide range of acquired and inherited disorders. 

January 10, 2014, Joseph G. Hacia, PhD - Associate Professor, Keck School of Medicine, University of Southern California 
Developing Targeted Therapies for Peroxisome Biogenesis Disorders: Biomedical Informatics Challenges and Opportunities 
Abstract: Peroxisome biogenesis disorders are recessive neurological disorders caused by defects in PEX genes required for the normal assembly of peroxisomes, intracellular organelles required for numerous metabolic processes essential for normal development and long-term health. Since the majority of patients have a degenerative form of disease that is compatible with survival through adulthood, more effective therapeutic interventions could have a major impact on their longevity and quality of life. Here, I will discuss multidisciplinary collaborative initiatives to discover and implement therapies that are addresses the molecular basis for disease. These include on-going high content drug screens that pose computational challenges for the processing of large libraries of cell images. I will also describe our experiences developing induced pluripotent stem cell models of disease for the next generation of quantitative high content drug screening. Finally, I will discuss other computational challenges and opportunities emerging for our efforts to develop stem cell transplant and gene therapies for peroxisome biogenesis disorders.