ISMB/ECCB 2013 Workshops
Heritability Analysis and Genetic Trait Prediction
Organizer(s):
Christoph Lippert is a researcher at Microsoft, Los Angeles. He is developing statistical and computational methodology for problems in genetics and genomics, with a primary focus on mixed models for genome-wide association studies. Before joining Microsoft he has been performing research at the Max Planck Institute for Develpmental Biology and the Max Planck Institute for Intelligent Systems in Tuebingen, Germany.
Bjarni J Vilhjalmsson is a postdoctoral fellow in the lab og Prof. Alkes Price at Harvard School of Public Health and the Broad Institute. He obtained his PhD under the guidance of Prof. Magnus Nordborg at the University of Southern California in 2011, where he developed mixed model methods for association mapping and heritability analysis. While at Harvard, Bjarni has shifted his focus to polygenic risk prediction with a focus on developing efficient methods for applications large sample size.
Room: Roof Garden
Genome-wide association studies involving tens of thousands of individuals have in recent years emerged as a powerful strategy for identifying genetic variants associated with common human diseases and complex traits. One remaining problem is that the genetic variants identified to date generally explain only a small fraction of the heritable trait variability as estimated from family studies, giving rise to the “missing heritability problem” (Maher, 2008 Nature). By estimating the heritability from genome-wide single nucleotide polymorphisms, Visscher and colleagues convincingly argued that most common diseases have a highly polygenic genetic architecture (Yang et al., 2010 Nat Genet; Yang et al., 2011 EJHG; Stahl et al., 2012 Nat Genet). This insight has important ramifications when developing methods for heritability estimation, and trait prediction. The highly polygenic nature of common diseases suggests that very large sample sizes are needed to accurately train models involving a large number of predictive markers, which further requires development of computationally efficient algorithms.
- 2001 - 2009 Professor of Statistical Genetics at Imperial College, London
- 1997 - 2001 Professor of Applied Statistics at the University of Reading
- 1989 - 1996 Lecturer and Senior Lecturer in Probability and Statistics, Queen Mary & Westfield College, U London
- 1988 - 1989 Junior Lecturer in Mathematical Statistics, University of Oxford, and Lecturer in Mathematics, Lady Margaret Hall, Oxford
- 1985 - 1988 D.Phil. in Mathematics, Trinity College, University of Oxford
Measuring the extent to which the genetic similarity of pairs of individuals can explain their phenotypic similarity has been at the heart of quantitative genetics for over a century. Until recently there was in effect only one definition of kinship, based on expected genome sharing computed from known pedigrees. But pedigrees only specify expected genome sharing, whereas nowadays genome-wide SNPs allow genetic similarity (kinship) between individuals to be measured directly. That there are many different kinship measures available has not been well recognised in the genetics literature, and this has contributed to confusion about the interpretation of modern heritability analyses. It turns out that defining kinship in a principled and useful way is surprisingly difficult, and it may now be time to abandon the idea that there is a "standard" single-number, or even low-dimensional, measure of relatedness, and hence that there is a single value for the heritability of a trait. This new flexibility allows traditional notions of heritability to be extended in many different ways, providing new tools for dissecting the genomic architecture of complex traits. I review these ideas and present some new approaches to genetic analyses based on SNP-based kinships. This is joint work with postdoc Doug Speed, who is funded by the UK Medical Research Council.
Resolving missing heritability, the difference between phenotypic variance explained by associated SNPs and estimates of narrow-sense heritability (h2), will inform strategies for disease mapping and prediction of complex traits. We develop a novel approach to estimating h2 based on sharing of local ancestry segments between pairs of unrelated individuals in an admixed population. Unlike recent approaches for estimating the heritability explained by genotyped markers (h2g) [Yang et al. 2010], our approach captures the total h2, because local ancestry estimated from genotyping array data captures the effects of all variants. Our approach uses only unrelated individuals, and is thus not susceptible to biases caused by epistatic interactions or shared environment. Theory and simulations show that the variance explained by local ancestry (h2γ) is related to h2, Fst, and genome-wide ancestry proportion (θ): h2γ = h2*2*Fst*θ*(1-θ). We apply our method to 5,040 African Americans and estimate the h2 for HDL cholesterol (0.39±0.11), LDL cholesterol (0.18±0.09), and height (0.55±0.13). These h2 estimates were higher than estimates of h2g from the same data: 0.22±0.07, 0.16±0.07 and 0.31±0.07. The difference between h2 and h2g suggests that rare variants contribute substantial missing heritability.
Dr. Eric Xing is an associate professor in the School of Computer Science at Carnegie Mellon University. His principal research interests lie in the development of machine learning and statistical methodology; especially for solving problems involving automated learning, reasoning, and decision-making in high-dimensional and dynamic possible worlds; and for building quantitative models and predictive understandings of biological systems. Professor Xing received a Ph.D. in Molecular Biology from Rutgers University, and another Ph.D. in Computer Science from UC Berkeley. His current work involves, 1) foundations of statistical learning, including theory and algorithms for estimating time/space varying-coefficient models, sparse structured input/output models, and nonparametric Bayesian models; 2) computational and statistical analysis of gene regulation, genetic variation, and disease associations; and 3) application of statistical learning in social networks, data mining, vision. Professor Xing has published over 150 peer-reviewed papers, and is an associate editor of the Journal of the American Statistical Association, Annals of Applied Statistics, the IEEE Transactions of Pattern Analysis and Machine Intelligence, the PLoS Journal of Computational Biology, and an Action Editor of the Machine Learning journal. He is a recipient of the NSF Career Award, the Alfred P. Sloan Research Fellowship in Computer Science, the United States Air Force Young Investigator Award, and the IBM Open Collaborative Research Faculty Award.
Genome-wide association (GWA) mapping have recently become a popular approach for identifying genetic loci that are responsible for increased disease susceptibility. In the presence of rich side-information about the structure of the data, such as population structures revealed by ancestral inference and heredity analysis, genome structures exposed in linkage disequilibrium studies, trait structures captured by expression networks or phenotype clusters, to name a few, traditional approaches such as p-value based pair-wise SNP-trait association tests, PCA, or lasso have difficulties in admitting (the constraints induced by) and exploiting (the benefits offered by) such information. In this talk, I will present a class of new models, algorithms, and theories that go beyond the traditional approach and enable effective use of structural information for GWA mapping in large-scale and high-dimensional setting given whole genome and phenome data. This approach builds on the sparse structured regression method in statistics, enjoying strong statistical guarantee and scalability, and can be flexibly configured to handle different structural information. I will demonstrate application of this approach to a number of complex GWA scenarios, including associations to trait networks or cluster-tree, to dynamic traits, under admixed populations, and with epistatic effects.
Oliver Stegle is a research group leader at the EMBl-European Bioinformatics Institute in Hinxton near Cambridge. He received his PhD from the University of Cambridge while working with David MacKay on machine learning and computational biology. He did his postdoctoral work at the Max Planck campus Tuebingen, working on genetic association mapping and statistical genomics methodology applied to plant systems and medical questions. He has broad interests in the genetic regulation of phenotype, molecular traits and how heterogeneous data sources can be effectively integrated.
Many phenotypic traits of interest are heritable and vary as a function of genetic and external factors. Association-based testing, heritability analysis and prediction of these phenotypes have yielded tremendous insights into the genetic mechanisms that cause phenotypic differences between individuals in a population. Most recent studies have begun to complement genotype and phenotype data with readouts from high-throughput molecular profiling techniques, such as RNA-Seq based gene expression profiling. These data reveal the genetic regulation of quantitative traits at an intermediate layer between genotype and phenotypes at an organismal level. In this talk, I will discuss how heritability analysis and phenotype prediction with BLUP-like methods can be adapted for molecular traits. To maximize predictive power, it is helpful to exploit existing prior knowledge about the genetic architecture and loci with known regulatory potential on individual genes. We propose a model that naturally incorporates such information and handles correlated molecular traits as they occur due to alternative RNA processing or molecular profiling in multiple related environments.
Bioinformatics Cores Workshop
Organizer(s):
David Sexton is the senior scientific project manager for the Novartis Institutes of Biomedical Research genome initiative. David was formerly the assistant director of the sequence analysis group at Baylor College of Medicine and Director of the computational genomics core at Vanderbilt University. He completed his masters in computer information systems and undergraduate degree in molecular and cellular biology from the University of Arizona.
Simon came to bioinformatics after training as a molecular biologist. He originally did a first degree in microbiology at Warwick University before moving on to do a PhD in molecular evolution at Newcastle University.
Simon always had a keen interest in the application of computing to biology and following his PhD he moved into a full time bioinformatics position. He initially worked at the BBSRC Bioscience IT Services where he was a protein analysis specialist and was responsible for providing training in bioinformatics to the various BBSRC Institutes.
Brent Richter is the Director of Enterprise Research Infrastructure and Services (ERIS) at Partners HealthCare, the corporation founded by Massachusetts General Hospital and Brigham and Women's Hospital in Boston MA. ERIS supports the IT and informatics needs for the nearly 8000 biomedical and basic research investigators that comprise the Partners academic medical centers.
Brent began his career at Harvard University developing the sequencing methodologies for the Human Genome project and subsequently applying the technologies and developing bioinformatics tools for the field of Population Genetics and Evolution. Over the last 12 years, he has moved to supporting investigators and physicians with HPC, scientific computing and research information services.
Matt Eldridge is the Head of the Bioinformatics Core at the Cancer Research UK Cambridge Institute (CRUK-CI) in the University of Cambridge.
The facility provides a bioinformatics analysis and statistics consulting service and develops software and data processing pipelines to support high-throughput technologies, especially sequencing.
Having originally trained in Computational Chemistry at the University of Oxford before moving into Bioinformatics at the European Bioinformatics Institute, Matt has over 10 years commercial experience in software development and delivery of enterprise data management systems in the pharmaceutical and biotechnology sector.
Hans-Rudolf Hotz, PhD, runs the Bioinformatics Helpdesk at the Friedrich Miescher Institute for Biomedical Research in Basel, Switzerland. Originally trained as a Molecular Biologist, he got his Bioinformatics education from the University of Manchester, UK. He was working during 8 years at the Wellcome Trust Sanger Institute in Hinxton (Cambridge), UK, first as a member of the web team, followed by head of the Core Bioinformatics team. In August 2008 he took over his current position.
Room: Roof Garden
The Bioinfo-core group is an international organization that brings together computational biologists working in and running bioinformatics core facilities. The environment in which core facilities operate means that they view the field from a perspective which requires them to weigh the scientific merits of analyses against factors such as ease of use, ease of interpretation, ability to scale, and robustness. Their exposure to varied datasets from many different sources allows them to critically evaluate software and methods from a viewpoint not normally available to research groups. They also have specific interests relating to the running of large multi-user facilities that the workshop could address. It is hoped that the unusual perspective of the group again justifies the inclusion of a workshop where both the scientific and technical aspects of undertaking bioinformatics in a core facility could be explored.
Topic1:Integrative analysis of large scale data
Topic2:Tracking and reporting the analysis of biological data
Mikhail graduated from Department of Biology at Moscow State University in 2002 and moved to London to do a PhD with Amanda Fisher and Matthias Merkenschlager at MRC Clinical Sciences Centre (Imperial College). His PhD projects investigated chromatin events underlying cell differentiation, and about half of his work was ‘wet-lab’ and half computational.
Mikhail then joined EMBL on an Interdisciplinary Fellowship to work with Ewan Birney (EMBL-EBI, near Cambridge) and Eileen Furlong (EMBL Heidelberg) on the regulatory organization of cardiac enhancers. At EMBL, he also contributed to the ENCODE project, studying the population genetics of transcription factor binding sites. In addition, Mikhail has been involved in the establishment of a population genomics resource in medaka fish that will become available in 2014.
In November 2012, Mikhail started a regulatory genomics group at the Babraham Institute as part of the Nuclear Dynamics ISP.
Suraj Menon graduated from the University of Mumbai with a BSc in Microbiology and Biochemistry. He then went on to complete an M.Res degree in Bioinformatics at the University of Glasgow. After working as a Bioinformatician at the Wellcome Centre for Molecular Parasitology, Suraj moved to Cardiff to pursue a PhD under David Kipling at Cardiff University. During this time he primarily worked in integration of microarray datasets and novel tools for unsupervised pathway analysis. Suraj is currently working as Senior Bioinformatics Analyst at the Bioinformatics Core of the Cancer Research UK Cambridge Institute. He specialises in analysis of high throughput data from various platforms and technologies including microarrays, next generation sequencing (ChIP-seq, RNA-seq) and proteomics (Mass spectrometry). His interests include pathway and network analyses, and integration of datasets across platforms and technologies.
This session will look at the increasing scale of the studies required to obtain a complete understanding of a biological system. The session will have two main focuses:
-
The integration of public datasets in the analysis of novel local data
-
The analysis of studies composed of multiple different data types
In the first part we will look at how relating results obtained from novel data to existing public datasets such as those developed by the Encode project, or those present in the major public data repositories can help with data interpretation. We will discuss the limitations of such analyses and cover the options we have for the types of data to use and the degree to which we trust existing data processing and analysis when integrating with our own data. We will also look at the tools available to make such converged public private projects as simple as possible. We can also discuss how to handle the management of large amounts of public data within an existing IT infrastructure to avoid repeated downloads or data duplication.
In the second part we will look at an increasing trend to use multiple different data types to profile a biological system. Already studies are being produced which incorporate RNA-Seq, ChIP-Seq of several factors, Bisulphite-seq and maybe mass spec and imaging data. These studies will become more common in future as will studies mixing in public datasets of different types. In this part of the session we will look at the strategies we can employ to identify complex signals across these types of data, the sorts of tools we can use to both visualise and analyse these types of experiment and discuss experiences people have had and share any recommendations.
Simon came to bioinformatics after training as a molecular biologist. He originally did a first degree in microbiology at Warwick University before moving on to do a PhD in molecular evolution at Newcastle University.
Simon always had a keen interest in the application of computing to biology and following his PhD he moved into a full time bioinformatics position. He initially worked at the BBSRC Bioscience IT Services where he was a protein analysis specialist and was responsible for providing training in bioinformatics to the various BBSRC Institutes.
David Sexton is the senior scientific project manager for the Novartis Institutes of Biomedical Research genome initiative. David was formerly the assistant director of the sequence analysis group at Baylor College of Medicine and Director of the computational genomics core at Vanderbilt University. He completed his masters in computer information systems and undergraduate degree in molecular and cellular biology from the University of Arizona.
Dr. Cavalcoli has broad research experience beginning with his PhD and post-doctoral research in molecular biology and virology. He received his PhD in 1993 from LSU, Baton Rouge in Virology and did a post-doc at University of Pittsburgh. In 1996, he made a transition to bioinformatics during a second post-doc at Parke-Davis in Ann Arbor. He has 12 years of experience with development and application of bioinformatics to drug discovery / industry research including mouse and human genetics, dyslipidemia, diabetes, cardiovascular disease, and schizophrenia. He has experience with proteomic informatics including development of relational databases and analysis of peptide fragment annotation via MALDI Mass Spectrometry. In addition, Dr. Cavalcoli has experience with analysis and interpretation of results from Affymetrix gene expression profiling.
Kim-Ahn undertakes a wide range of valuable collaborative and research opportunities in both statistics and molecular biology. Her research interests are multidisciplinary, focusing on mathematical statistics characterisation of molecular biological systems, and she is interested in developing sound statistical frameworks applied to addressing new biological questions arising from these frontier molecular technologies.
Kim Ahn is currently developing statistical approaches to analyse high throughput data. Her research focuses on ‘variable selection’, i.e. how to identify relevant information in such large data sets, using both classification and exploratory approaches.
She is an associate lecturer in the School of Chemistry and Molecular Biosciences at the University of Queensland, and teaches in the University’s Bioinformatics Masters program.
A central tenet of science is that all results should be described in sufficient detail to allow them to be reproduced by others in the field. Computational analyses are forming an ever larger proportion of major papers and yet it is still often difficult to reproduce exactly the analyses described in many papers. On a more practical level a core facility has to frequently report the results of an analysis back to a scientist and to do this effectively they need systems to keep track of the analysis they have done (which may involve many blind alleys before hitting the final result), and to present this in an understandable and yet robust way.
This session will look at how different core facilities keep track of the work they do. It will look at any tools or systems people use, and the level of detail they use in recording the analysis. We will discuss the potential conflict between recording all steps in an analysis rigorously and the overhead this imposes on the speed at which different analyses can be tried.
Some groups are now producing completely automated records of their analysis in a format which allows them to be re-run on other sites. We can discuss how useful this might be in the field, both for reproducing existing analyses and for constructing new pipelines based on previous results. We can also discuss the conflict between having full details of an analysis suitable for a computer to reconstruct it and producing clear reports which explain the process undertaken in a simple way to scientists.
Hans-Rudolf Hotz, PhD, runs the Bioinformatics Helpdesk at the Friedrich Miescher Institute for Biomedical Research in Basel, Switzerland. Originally trained as a Molecular Biologist, he got his Bioinformatics from the University of Manchester, UK. He was working during 8 years at the Wellcome Trust Sanger Institute in Hinxton (Cambridge), UK, fFirst as a member of the web team, followed by head of the Core Bioinformatics team. In August 2008 he took over his current position.
Matt Eldridge is the Head of the Bioinformatics Core at the Cancer Research UK Cambridge Institute (CRUK-CI) in the University of Cambridge.
The facility provides a bioinformatics analysis and statistics consulting service and develops software and data processing pipelines to support high-throughput technologies, especially sequencing.
Having originally trained in Computational Chemistry at the University of Oxford before moving into Bioinformatics at the European Bioinformatics Institute, Matt has over 10 years commercial experience in software development and delivery of enterprise data management systems in the pharmaceutical and biotechnology sector.
Open discussion of issues facing bioinfomatics cores
What Bioinformaticians need to know about digital publishing beyond the PDF
Organizer(s):
Dr. Marco Roos is Research Scientist in Human Genetics at Leiden University Medical Center, and co-director of the Biosemantics Group. Marco also holds a position at the Informatics Institute of the University of Amsterdam, and leads the Interoperability Task Force of the Netherlands Bioinformatics Centre (NBIC). His research interests are e-Science (Semantic Web and workflows) and the application of knowledge discovery techniques for elucidating molecular mechanisms in the cell, in particular the role of epigenetics.
Oscar Corcho is an Associate Professor at Departamento de Inteligencia Artificial (Facultad de Informática , Universidad Politécnica de Madrid) , and he belongs to the Ontology Engineering Group. His research activities are focused on Semantic e-Science and Real World Internet, although he also works in the more general areas of Semantic Web and Ontological Engineering.
After training in Biochemistry from Imperial College, and a PhD on the Molecular Pathology of Ocular Melanoma at the Royal London Hospital his research mainly focused on cancer cell and molecular biology. After postdocs at the WHO International Agency for Research in Cancer in Lyon and Queen Mary University of London, he was senior scientific editor for the BMC Genomics and Bioinformatics journals at BMC before moving in 2010 to Shenzhen/Hong Kong to set up the GigaScience journal, database and integrated data analysis platform for the BGI.
Carole Goble is a full professor in Computer Science at the University of Manchester, UK. She has an international reputation in Semantic Web, Distributed computing, and Social Computing for scientific collaboration. She is a member of the Software Sustainability Institute UK. She directs the myGrid project, known for Taverna, myExperiment, Biocatalogue, and SEEK. In 2008 Carole was awarded the inaugural Microsoft Jim Gray award for outstanding contributions to e-Science. In 2010 she was elected a Fellow of the Royal Academy of Engineering. In 2012 she was nominated for the Benjamin Franklin award for open science in Biology.
Barend Mons is professor in Biosemantics at the Department of Human Genetics at the Leiden University Medical Centre, and Scientific Director of the Netherlands Bioinformatics Centre (NBIC). He is the initiator of ConceptWiki.org, an inventor of Knowlet technology for knowledge discovery, and a driving force behind the Concept Web Alliance. He also co-founded the companies Collexis and Knewco. Mons published over 70 peer reviewed articles and holds three patents in semantic technology.
Jun Zhao is a Research Officer in the Image Bioinformatics Research Group of the Department of Zoology, University of Oxford. She is interested in applying Semantic Web technologies for publishing research data (including images) to the Web. She is an active member in the Linked Data research community.
Dr. Erik Schultes is Research Scientist in Human Genetics at Leiden University Medical Center, and co-director of the Biosemantics Group. His research centers on in silico knowledge discovery, especially the development of publication formats that incentivize data interoperability and the application of complexity theory to automated semantic reasoning.
Room: Roof Garden
Data, software and numerous modes of written communication in the life sciences are rapidly increasing in volume, frequency, and sophistication. In addition to traditional research articles being on-line, there is an explosion of alternative publication formats explicitly created for digital environments (microattribution, Tiny ToCS, nanopublications, Research Objects, ISA). Live resources, such as databases, source code repositories, web services and workflows are rapidly becoming essential tools for the life sciences and blur the line between data-production and the communication of results. Furthermore, innovative platforms for crowd-sourced knowledge discovery are emerging where a vast landscape of findings and hypotheses can be made public in real-time data streams such as twitter, blogs, wikis, etcetera. Tracking trends and capturing emerging knowledge in this ecosystem is becoming a vital area of commercial research and development.
The aim of this workshop is to inform participants of ISMB/ECCB of changes and new opportunities in scientific communication. It addresses guidelines and tools that enable contributions to be exposed in machine-readable formats, ready for evaluation, interoperation, reuse, and attribution. We discuss how this creates an incentive for researchers, data curators, and even non-researchers (e.g. patients), to publish valuable data that bioinformaticians can subsequently use for new analyses.
Associate Vice Chancellor for Innovation and Professor of Pharmacology at the University of California San Diego, Founding Editor in Chief of PLOS Computational Biology, Associate Director of the RCSB Protein Data Bank, Co-founder of SciVee Inc. and a founding member of The Future of Research Communicatione and e-Scholarship (FORCE11; www.force11.org)
After a short introduction by Prof. Carole Goble, Prof. Phil Bourne will cover broad issues and initiatives in the digital publication of algorithms, text mining results, machine learning, database and other bioinformatics resources. He will also emphasize how new publication modes can promote the free dissemination and widespread integration of scientific knowledge.
Rebecca Lawrence is Managing Director of F1000 Research Ltd. She launched both of the organisations' major products: the open access repository of posters and slides F1000Posters, and the open access life sciences journal F1000Research that redefines scientific publishing through immediate publication, transparent formal peer review post-publication, and mandatory data deposition. She has previously worked for other publishers including Elsevier for 7 years running the Drug Discovery Group, and holds a degree in Pharmacy and a PhD in Pharmacology.
Rebecca Lawrence, Managing Director of F1000 Research Ltd, will present the novel publishing opportunities that the open access journal F1000Research can offer to bioinformatics researchers due to its unusual model of immediate publication, transparent formal peer review post-publication, and mandatory data deposition. F1000Research is collaborating with the digital library community on employing new standards for capturing structured research methods and results. She will speak about the impact of new digital publication methods, in particular for results that do not align well with traditional publishing models, and the challenges that brings in reviewing methods and data.
After training in Biochemistry from Imperial College, and a PhD on the Molecular Pathology of Ocular Melanoma at the Royal London Hospital his research mainly focused on cancer cell and molecular biology. After postdocs at the WHO International Agency for Research in Cancer in Lyon and Queen Mary University of London, he was senior scientific editor for the BMC Genomics and Bioinformatics journals at BMC before moving in 2010 to Shenzhen/Hong Kong to set up the GigaScience journal, database and integrated data analysis platform for the BGI.
Dr. Scott Edmunds will describe how the open-access, open-data journal GigaScience is attempting to revolutionize large-scale biological data dissemination, organization and re-use. Utilizing the BGI’s experience and data handling infrastructure, Scott will describe and provide examples of their novel publication format: linking standard manuscript publication with an integrated database that hosts all associated data and provides data analysis tools and computing resources. Making publishing more transparent and open, authors are credited for making all of the supporting data, workflows and methods available, promoting reproducibility, and with the ultimate goal of producing executable papers.
Barend Mons is professor in Biosemantics at the Department of Human Genetics at the Leiden University Medical Centre, and Scientific Director of the Netherlands Bioinformatics Centre (NBIC). He is the initiator of ConceptWiki.org, an inventor of Knowlet technology for knowledge discovery, and a driving force behind the Concept Web Alliance. He also co-founded the companies Collexis and Knewco. Mons published over 70 peer reviewed articles and holds three patents in semantic technology.
Niklas Blomberg is Founding Director of ELIXIR, the European infrastructure for bioinformatics and life-science data, based in Hinxton, UK. He holds a BSc in Chemistry (Göteborg) and a PhD from EMBL, Heidelberg. In 1999, he joined AstraZeneca working on topics such as fragment-based lead generation and compound screening collection enhancements. As Associate Director of Computational Chemistry, he led the global cheminformatics group. In 2011-2013, he built a new team to support Respiratory, Inflammation and Autoimmune research. He is an industry advisor in national eScience initiatives, until recently chaired BILS (Swedish ELIXIR node), and has co-chaired the IMI Open PHACTS initative.
Prof. Lawrence Hunter is the Director of the University of Colorado's Computational Bioscience Program. Dr. Hunter is one of the founders of bioinformatics, serving as the first President of the International Society for Computational Biology and contributing to the creation of the ISMB, PSB, Rocky and VizBi conferences. He has published more than 100 scientific papers, holds two patents and has been elected a fellow of both the ISCB and the American College of Medical Informatics. His research focus is the integration of natural language processing, knowledge representation, machine learning and visualization techniques to help interpret genome-scale data.
Carole Goble is a full professor in Computer Science at the University of Manchester, UK. She has an international reputation in Semantic Web, Distributed computing, and Social Computing for scientific collaboration. She is a member of the Software Sustainability Institute UK. She directs the myGrid project, known for Taverna, myExperiment, Biocatalogue, and SEEK. In 2008 Carole was awarded the inaugural Microsoft Jim Gray award for outstanding contributions to e-Science. In 2010 she was elected a Fellow of the Royal Academy of Engineering. In 2012 she was nominated for the Benjamin Franklin award for open science in Biology.
A panel of leaders in the field of digital publishing will field audience questions. This session aims to provide clarity about digital publishing opportunities for bioinformaticians, and help them keep on top of new mechanisms in research communication. The panel consists of our speakers augmented by Dr. Niklas Blomberg (director of Elixir), Prof. Carole Goble (University of Manchester, PI of myGrid, myExperiment and BioCatalogue), and Prof. Larry Hunter (University of Colorado, ISMB founder and researcher of knowledge-based techniques). Prof. Barend Mons will chair, after giving a short overview of digital publishing models and how they relate.
Workshop on Education in Bioinformatics (WEB 2013)
Organizer(s):
Michelle Brazas, PhD is the Coordinator and lead of the Canadian Bioinformatics Workshops (bioinformatics.ca) and Manager of Bioinformatics Education at the Ontario Institute for Cancer Research. She also teaches bioinformatics in Nigeria as part of her outreach programs.
Patricia Palagi, PhD is the Coordinator of the SIB Education Activities and coordinator of the Bioinformatics track of the Master’s in Biology common to the Universities of Geneva and Lausanne. She teaches bioinformatics for proteomics and programming languages, and organises a variety of bioinformatics training, workshops and events for a wide audience.
Vicky Schneider, PhD is part of the Senior Management Team at TGAC, where she is Head of the Training and Outreach Team. She also facilitates and delivers actual training for a variety of audiences on bioinformatics and related subjects.
Celia van Gelder, PhD is project leader education of the Netherlands Bioinformatics Centre (NBIC, www.nbic.nl) and coordinator bioinformatics education at the Radboud University Nijmegen. She also teaches bioinformatics to undergraduates and life science researchers. She is treasurer of GOBLET (mygoblet.org – Global Organisation for Bioinformatics Learning Education and Training).
Room: Roof Garden
Traditionally bioinformatics tools and training programs have focused on the life science audience. Though heterogeneous, the needs of the life science audience are at least fairly well understood. Driven by the impact of technology in diverse areas, bioinformatics is becoming increasingly interdisciplinary, and in parallel so too are the audiences seeking bioinformatics training. Audiences as disparate as physicians and lawyers, industry and even the general public, previously without real need of bioinformatics skills, are now pursuing an understanding and skill set in bioinformatics.
These new audiences represent an exciting new challenge for bioinformatics training programs. Each audience brings a specific training need and a mix of learning styles. Through a series of presentations and discussions, this workshop aims to bring awareness to the bioinformatics training needs of these new audiences and to provide a platform for sharing training strategies and challenges experienced by various programs reaching out to these diverse audiences.
This workshop will consist of presentations on the topics of 1) how a medically trained audience thinks and learns about bioinformatics and the opportunities for new bioinformatics tools for the clinical setting; and 2) opportunities and strategies for providing bioinformatics training to an engaged, savvy public.
Russ Altman is a Professor at Stanford University Medical School, where he is chair of the department of Bioengineering and director of the Biomedical Informatics program. He is a past president and founding member of ISCB. His research interests are in the application of computing technology to basic problems in molecular biology of relevance to medicine. He has a particular interest in informatics methods to support pharmacogenomics, the study of how genetic variation impacts drug response phenotypes. Dr Altman has a long experience in teaching and training bioinformatics. He will speak about "Biomedical informatics training in the era of translational medicine".
The Stanford Biomedical Informatics (BMI) program has provided training in informatics applied to biology and medicine since 1982, and has more than 100 PhD and MS graduates. The program has evolved to support a broad range of biomedical informatics interests including: bioinformatics, clinical informatics, imaging informatics. Most recently "translational bioinformatics" has become an area of great interest among students. In this talk, I will briefly summarize the key features of the Stanford BMI curriculum and student program, and the key core classes that are taught by our faculty. I will discuss the evolving interests and backgrounds of applicants to our program. We have built a distance MS degree and are experimenting with other remote learning technologies. Finally, I will review the outcomes for our students, and summarize where our alums have landed and what they are doing.
Donna Slonim is an Associate Professor at Tufts University, with appointments in Computer Science, Pathology, and Genetics. She has worked on the Human Genome Project, conducted pharmacogenomics research at Wyeth Pharmaceuticals, and developed gene expression data analysis methods. Her current research focuses on using genomics to identify new diagnostic and treatment approaches for developmental disorders.
Donna Slonim is a translational researcher who has taught bioinformatics to present and future clinicians in several contexts. She will speak about successful translational collaborations and about bioinformatics training for doctors. In particular, she will describe her own experiences in designing bioinformatics courses for undergraduates interested in medicine, and she will discuss the changing needs for pre-medical quantitative education at the undergraduate level.
Ms. Hienke Sminia (MSc) is the Education Officer at NBIC and is responsible for the high school projects within NBIC education. This programme also organizes activities and develops education material for the general public.
Hienke Sminia will share her experiences with activities and exhibits for the general public (from ages 5 to 99). She focuses on the content suitable for the different age-groups and settings, as well as opportunities to engage this audience in bioinformatics.
Dr. Winston Hide is an Associate Professor of Bioinformatics and Computational Biology at the Harvard School of Public Health, and is affiliated with the South African National Bioinformatics Institute.
Winston Hide maintains a lively blog on genomics and bioinformatics related issues for the public audience. Dr. Hide will speak on strategies used to engage the public in genomics and informatics understanding.
The “How To Guide” for Establishing a Successful Bioinformatics Network
Organizer(s):
Aidan is a Senior Computational Biologist at the European Molecular Biology Laboratory (EMBL) in Heidelberg, Germany www.embl.de. Within this role, he is project manager of an internal (to EMBL) professional bioinformatics network, and is an organizer of the Heidelberg Unseminars in Bioinformatics local professional bioinformatics network. He regularly organizes and teaches in international bioinformatics training courses focused on computational molecular evolution, and protein sequence analysis.
Michelle is the Coordinator of the Canadian Bioinformatics Workshops (bioinformatics.ca) and Manager of Bioinformatics Education at the Ontario Institute for Cancer Research. She also teaches bioinformatics in Nigeria as part of her outreach programs, and has previously organized workshops at two previous ISCB meetings i.e. Workshop on Education in Bioinformatics (WEB) 2011 and 2012.
Manuel is Project Leader for Plant and Animal Genomes at The Genome Analysis Centre, UK. Manuel is known in the bioinformatics community as the founder of ISCB’s Student Council, the leading bioinformatics student organization globally. He also helped establish the ISCB Africa ASBCB conference, the main Bioinformatics event in the African continent. Manuel is director of Itico, a non-profit organization supporting event sharing among scientific organizations. Manuel is also involved in the GOBLET organization and ISCB’s Junior PI Initiative.
Room: Roof Garden
Collaborative networks are increasingly common in bioinformatics (i.e. groups of people working together to achieve bioinformatics-related goals), fuelled by rapid growth in data production and Internet bandwidth. Key factors determining their successful development, however, remain somewhat obscure.
Face-to-face contact between professionals with common interests is often crucial for successfully establishing and maintaining professional networks. For bioinformaticians, ISMB/ECCB 2013 is thus an ideal environment for fostering the development of such networks. This workshop aims at bringing together key members of successful networks that are playing important roles shaping the field of bioinformatics. The workshop is specifically designed to engage all participants, not just invited speakers, in identifying together key factors involved in establishing and maintaining successful, sustainable bioinformatics networks.
This should help participants in assessing the viability of prospective bioinformatics projects or collaborations relying on networks of professionals, and by providing insight in how to contribute directly to the success of networks they already belong to, or those they are instrumental in establishing.
The results of the workshop will be published, to make them available to the global bioinformatics community. The twitter hastag for the workshop is #ismbnetworking; you can also join the LinkedIn group for people interested in the workshop.
Reinhard Schneider, current treasurer of the ISCB, is actively involved in various internationally-recognized networks and advisory boards. These include, amongst others, the Garuda Alliance, an international collaboration to provide a common platform for integration of systems biology bioinformatics tools, EMBnet, scientific advisory boards and smaller-scale projects such as internal interdepartmental networks of bioinformatics researchers. Dr Schneider is currently the Head of Bioinformatics Core Facility at the Luxembourg Centre for Systems Biology.
Jon works with Prof. Rebecca Wade at the Heidelberg Institute for Theoretical Studies, where he is a member of the Virtual Liver Network. During his PhD studies in structural bioinformatics of protein-protein interaction inhibitors, with Dr. Richard Jackson at the University of Leeds, he won a CCPB scholarship to work with Prof. Michael Shirts at the University of Virginia. He is a founding member of Heidelberg Unseminars in Bioinformatics, where he coordinated the discussion and collaborative writing by HUB participants of 'Biggest Challenges in Bioinformatics', a short article recently published in EMBO reports. Find Jon on twitter @JonathanCFuller.
Presentation: Why Professional Bioinformatics Networks are Important
To put the importance of the topic of the workshop in context, the workshop begins with a short description of how networks can benefit and support bioinformatics activities. This will be presented by Reinhard Schneider, current treasurer of the ISCB, who participates in many international networks.
Small-Group Discussion and Networking Activity
Many workshop participants will be involved in one or more professional bioinformatics networks. A key aim of the workshop is to engage with the knowledge and experience of all participants, not just invited speakers, in understanding what makes such networks successful i.e. efficiently achieves their goals, where participation provides significant benefits to the members of the network, and which are sustained as long as it is useful i.e. are continuing to achieve their goals.
To involve all participants in these discussions, most of this session will involve a networking activity, with all participants taking part in a set of parallel small-group discussions, to share together examples of successful networks in which they participate.
The networking session will be coordinated by Jonathon Fuller. a founding member of Heidelberg Unseminars in Bioinformatics (HUB) a participant-driven, networking-focused forum for people interested in bioinformatics.
Wolfgang Huber is a group leader and senior scientist at the EMBL in Heidelberg. He has a PhD in statistical physics from the University of Freiburg, and worked as a postdoctoral researcher at IBM Research, San Jose, California, at the DKFZ, Heidelberg, and as a research group leader at the EBI in Cambridge, UK. Wolfgang has been a long-term participant in the Bioconductor project, has written and is maintaining several software packages, and he is currently a member of Bioconductor's Advisory Board.
Jeremy Goecks is a Postdoctoral Researcher in the Department of Math & Computer Science and the Department of Biology at Emory University. He is also a lead member of the Galaxy project. Galaxy is a popular Web-based platform for doing accessible, reproducible, and collaborative computational biomedical research. Jeremy leads the development of Galaxy’s collaboration and visualization features and its cancer genomics pipelines. His publications have appeared in numerous biotechnology journals and top-tier computing conferences.
The second and third sessions of the workshop will include presentations describing four different successful professional bioinformatics networks, selected to showcase how different and diverse such networks can be.
Bioconductor is a highly successful open-source software project, based on the statistical programming language R. It involves many individual contributors from around the world, and has been running for more than 10 years. In addition to extensive virtual interaction, the project also involves regular face-to-face meetings and courses.
Galaxy is an open source, web-based platform for data-intensive biomedical research. Galaxy has several mechanisms for community based networking including mailing lists, a forum and code repository, as well as tools for easy sharing of workflows and results among project members
A/Prof Mulder heads the UCT Computational Biology Group at the University of Cape Town and is Coordinator of H3ABioNet, a Pan African bioinformatics network for H3Africa consisting of 34 institutions in Africa and the USA. After her PhD in Medical Microbiology, she spent 8 years at the European Bioinformatics Institute first as a curator and then as InterPro Team Leader. At UCT A/Prof Mulder works in the bioinformatics of microbial genomics, human genetics and infectious diseases, and provides Bioinformatics services for the local researchers. Through H3ABioNet, this service provision and training extends to Bioinformatics groups in Africa and H3Africa researchers.
Magali Michaut is a computational biologist working in the field of cancer genomics in the Computational Cancer Biology CCB) group of the Netherlands Cancer Institute. She has been involved in the ISCB Student Council since 2007, serving as the elected secretary in 2009 and actively taking part in the organisation of various symposia. She co-chaired the 6th Student Council Symposium in Boston and created the European Student Council Symposium in Ghent in 2010. Magali has been developing other local networks by creating Regional Student Groups (RSG) in France and in Europe.
The second and third sessions of the workshop will include presentations describing four different successful professional bioinformatics networks, selected to showcase how different and diverse such networks can be.
H3ABioNet aims to deliver a sustainable African bioinformatics network, providing user support, research, and training infrastructure for the pan-African H3Africa consortium. H3ABioNet is an example of a well-funded network with strong face-to-face and virtual components. Jointly funded by the NIH and the Wellcome Trust, a key aim of this project is to facilitate genomics research into common diseases having a major impact on African populations.
ISCB Student Council (SC) is a group led by students and for students. Its goal is to promote the development of the next generation of computational biologists, specifically through provision of scientific events, networking opportunities, soft-skills training, educational resources and career advice, while attempting to influence policy processes affecting science and education.
Regional Student Groups (RSGs) were first initiated by the ISCB Student Council in 2006, to fill its missions at the local level and address the needs of students in the field of computational biology in each region. Over the past few years, the student council has successfully established over 20 RSGs worldwide.
Francis Ouellette is a Senior Scientist, Associate Director of the Informatics and Biocomputing platform at the Ontario Institute for Cancer Research (OICR) as well as associate professor in the department of Cell and Systems Biology at the University of Toronto. He oversees the evaluation and development of pipelines for the analysis of NGS cancer data and also oversees the curation of the International Cancer Genome Consortium (icgc.org) data. Francis is also the scientific coordinator and instructor with the Canadian Bioinformatics Workshop series (CBW - http://bioinformatics.ca). His research interests include building and integrating biological databases, and the analysis of cancer genomes
Dr Blomberg holds a BSc in Chemistry from Göteborg University, Sweden, and a PhD from the European Molecular Biology Laboratory, Heidelberg, Germany. Before becoming Director of ELIXIR, the European infrastructure for biological information, he was Principal Scientist and Team Leader in Computational Chemistry and Computational Biology at AstraZeneca R&D Mölndal, Sweden. He has been an industry advisor in national eScience initiatives and an active participant in cross-industry research programmes.
Aleksandra Pawlik works at the Software Sustainability Institute supporting the development of research software communities in various disciplines. Aleksandra co-manages the Institute's Fellows' network and is responsible for outreach and seeking new opportunities for collaboration. Her work is also related with the myGrid project. She holds an undergraduate degree in Computer Science from Jagiellonian University and an MSc in Sociology from the University of Edinburgh. Outside of academia, Aleksandra has worked as a researcher for NHS Lothian projects and as a freelance IT consultant in the commercial sector.
The panel will include representatives from three additional successful bioinformatics networks. Panelists will present their experience on several differnet topics, after which discussion of these issues will be opened to the all participants. The networks represented on the panel will be:
- ELIXIR, www.elixir-europe.org, one of the largest and most intricate professional bioinformatics networks to date. The ELIXIR project aims to establish a set of sustainably-funded infrastructure for biological information within Europe.
- TorBug (Toronto), VanBug (Vancouver) and MonBug (Montreal), the Canadian bioinformatics user groups
- myGrid produce and use a suite of tools designed to “help e-Scientists get on with science and get on with scientists”. The tools support the creation of e-laboratories and have been used in domains as diverse as systems biology, social science, music, astronomy, multimedia and chemistry.
- Software Sustainability Institute Fellows' Network is a multidisciplinary group of researchers gathering intelligence about research and software from their respective disciplines, developing interests in the area of software sustainability, capacity building and policy development
Integrated Visualization and Computing for Biomedical Research
Organizer(s):
Nils Gehlenborg is working on data visualization and data exploration challenges related to largescale biological data sets and has been involved in the development of several data visualization and analysis systems. He is a co-founder of the IEEE Symposium on Biological Data Visualization (BioVis) and was a General Chair for the meeting in 2011 and 2012. He is also a co-founder of the VIZBI workshop series, co-organizer of panels, BoF meetings and workshops on biological data visualization at ISMB/ECCB 2009, 2010 and 2011 and at IEEE VisWeek 2010.
Jeremy Goecks leads the design and development of visual analysis tools for Galaxy, the popular opensource genomics platform. He has published on visual analysis tools and techniques for large biomedical data, with a focus on nextgeneration sequencing data and analyses. His publications have appeared in both computer science conferences (e.g. IEEE Symposium on Biological Data Visualization) and biomedical journals (e.g. Nature Biotechnology). He has been an invited speaker for past workshops at ISMB, ASHG, and other biomedical meetings, and he regularly participates in BoF meetings on biological data visualization.
Dave Clements coordinates training and outreach efforts for the Galaxy Project. He has led training and outreach activities for two large open source projects (Galaxy and GMOD) and has over five years of experience organizing workshops, including the "Genomics for Non-Model Organisms" workshop at ISMB/ECCB 2011, and workshops at Biocuration, Arthropod Genomics, and SMBE. He also organizes the annual Galaxy community meetings, and organized several GMOD summer schools and other events in the past. He has worked extensively with visualizing gene expression patterns (in situ, manually annotated, and 3D) and their relationship to anatomy ontologies.
Room: Roof Garden
Visualization has proven to be an essential tool as biological data sets have increased in size and complexity. A major trend in biological visualization is the integration of visualization with computing to create integrated analysis environments, where computational tools and visualization are used together for complex scientific investigations, such as hypothesis generation and evaluation. By combining complementary tools, such integrated environments have proven very useful for computational biology.
Realizing integrated visualization and computing environments for computational biology is challenging. First, software infrastructure is needed for implementing visualization tools that can support large and complex biological data sets and provide access to analysis tools. Additionally, there are often no visual encodings for biological data sets that integrate multiple data types across many experimental conditions or time points. This is exacerbated by the fact that training in data visualization and human-computer interaction techniques is typically not part of bioinformatics and computational biology curricula.
The workshop will address these challenges by highlighting successful approaches and outstanding challenges for creating integrated visualization and computing environments. Participants will leave the workshop equipped with an overview of the state-of-the-art, an understanding of the major pitfalls and pointers to additional resources.
Nuria Lopez-Bigas is a biologist with extensive expertise in computational genomics. Her lab is focused on the computational analysis of genomics data for the study of cancer. They have developed IntOGen, a resource for the analysis and integration of cancer genomics data (http://www.intogen.org), and new methodologies for the visualization and analysis of genomics and transcriptomics data (e.g. Gitools, OncodriveFM, Condel, transFIC, SLEA).
Large collections of datasets have become commonplace in biological research due to the increasingly pervasive use of high-throughput biology technologies (e.g., DNA sequencing). For visual analysis tools to be effective, approaches are needed for segmenting, aggregating, analyzing, and visualizing dataset collections. Effectively sorting and grouping datasets (e.g. by metadata such as tissue type or by features such as gene expression levels) often requires using computation for aggregation and visualization for evaluation of results. To visualize heterogeneous datasets, multiple visualization tools must often be integrated together or new, comprehensive visualizations developed. Both approaches are technically challenging. Finally, new computational strategies are needed so that visual analysis tools can produce results in seconds or minutes so that investigators can rapidly iterate on analysis results using an interactive interface.
Jan Aerts (http://homes.esat.kuleuven.be/~bioiuser/person.php?persid=473) is Assistant Professor in Bioinformatics at the University of Leuven. He received a PhD from Wageningen University (Netherlands) in 2005 on the Chicken Genome Sequencing Project. He subsequently worked as a postdoc at the Roslin Institute in Scotland (on the Cow Genome Sequencing Project), and as senior bioinformatician at the Wellcome Trust Sanger Institute in England. In 2010 he accepted a position in Leuven on biological data visualization. Prof Aerts' research interests focus on the use of visualization and visual analytics in genomics and (personalized) medicine.
Integrating visualization and analysis tools to create visual analysis applications requires creating an interface between the two components. Approaches that make it easier to combine visualizations and tools would lower the barrier for creating visual analysis applications and likely speed up the development of new applications. Common strategies that apply to different application domains (e.g., genomics, proteomics), different visualization types (e.g., genome browsers, heatmaps), and different analysis tools (e.g., command-line, Web-based) are most needed.
Carsten Görg is a faculty member in the Computational Bioscience Program in the University of Colorado Medical School. He received a PhD in computer science from Saarland University, Germany in 2005, on the layout of dynamic networks and then worked as a Postdoctoral Fellow in the Graphics, Visualization & Usability Center at the Georgia Institute of Technology before joining the University of Colorado. Dr Görg's research interests include visual analytics and information visualization with a focus on designing, developing, and evaluating visual analytics tools to support the analysis of biological and biomedical datasets.
Adopting a user-centered perspective for designing and evaluating visual analysis tools has many potential benefits, including better designed tools that can more effectively support the user’s analytical tasks. However, adopting a user-centered design approach also adds an additional layer of complexity to the design and evaluation of tools and it can be challenging for developers to acquire the necessary expertise in that area. Existing guidelines and methodologies can help guide developers through the essentials of user-centered design and tool evaluation. The analysis of complex datasets in the bioinformatics domain often cannot be completely automated but requires significant user interaction. Visual analysis tools can be very beneficial to support exactly these types of analyses.
The final session of the workshop will be a panel discussion with the invited speakers moderated by the organizers. This session will give workshop participants and speakers the opportunity to discuss cross-cutting issues and explore additional topics beyond the three main challenges covered by the talks.