Eric Karl Neumann

Professional Background Current Positions Director, Clinical Semantics Group MIT Fellow, Science Commons, CSAIL/MIT co-chair, W3C Healthcare and Life Sciences Interest Group Editorial Boards Frontiers in Neuroinformatics Drug Discovery Today Curriculum Vitae I am admittedly an evangelist for the Semantic Web, but I am also an implementor of it (one of many). Currently, I am director of the Clinical Semantics Group, a consulting firm specializing in semantic web solutions for pharmaceutical R&D. I have worked in the biotech and pharmaceutical world for the last 14 years, and in a few large corporations such as BBN and Sanofi-Aventis Pharmaceuticals. I have also helped launch three separate start-ups, and established novel approaches in informatics systems, knowledge management, and biomarkers applications. I helped create and over-saw BG-Medicine's informatics division, and co-founded Genstruct where I proposed their Knowledge Assembly strategy. Both of these companies have successful collaborations with pharmaceutical companies. I am also an MIT Fellow with Science Commons, working on the NeuroCommons Project. Throughout my career I have initiated and supported several consortia for the advancement of better scientific and informatics practices. It's a substantial activity that I believe is worth supporting in order to make real innovation possible. Some of the results of these efforts are being applied in several academic collaborations I am part of. BTW, in the spirit of taking one's own medicine, these pages have RDFa embedded in them (use a RDFa Bookmarklet to see and extract them).

Consortia / Projects

• W3C Healthcare and Life Sciences Interest Group
• BioPathways Consortium
• BioPAX
• Bio-Ontologies
• Drug Safety and Efficacy

---

I can be contacted via email.

About Myself

I am originally a neurobiologist who transitioned into informatics, providing yet another example that information drives biology. I worked at BBN for 7 years, and met many people who showed me the deeper relations between computers and humans. Outside of my work I enjoy art, nature, and literature, but have a personal passion for music-- when I can find the time. There is a set of natural relations between science and art, and this makes the work so much more enjoyable.

Back in 2004, a few of us were able to jump-start the Semantic Web initiative in the Life Sciences. I am now a co-chair for W3C's Semantic Web for Healthcare and Life Sciences Interest Group where we have been trying to engage the community in real informatics solutions based more on practical and attainable approaches. I have seen many efforts focus more on marketing, rather than advancing innovative and needed tools for improving human health. My hope is to make a real impact here.

I strongly believe that only through the use of explicit semantics can we effectively describe things like genes, proteins, compounds in familiar terms, and without artificial formalisms (e.g., database schemas). This has motivated me to try and apply emerging Semantic Web technologies, on interesting and challenging problems. Although I am a bit of an evangelist on this topic, I have actually worked on several examples involving pharmaceutical and clinical data, including a dashboard application called BioDash. I am currently working on a new technology called 'MagLev' that combines database integration and social web structures in a Semantic Web model.

Finally, I've really become excited about the Exhibit semantic viewing technology from MIT's SIMILE project -- you can see some interactive demos of it for viewing Clinical Trial data , Kinase Disease relations and Cancer Genome information here.

Sports and other Activities

• nordic and alpine skiing, volleyball, softball, fencing
• hiking, sailing, snorkeling, canoeing
• performing/improvising music, rock bands

My Interests

Music

As so many others did, I started out playing clarinet in the school band; I liked it and tried very hard to get the chops for playing classical repertoire, but I never could make the jump to jazz (which was way cooler), so I eventually lost interest. Wanting to be an aspiring 'rock star', I bought myself a bass guitar, picked up keyboard along the way, and had a lot of fun performing in bands. Now however, I have begun trying to create some of my own compositions, and have forced myself to get better at playing piano. Some of my compositions can be heard from my Music Page.

Recently, a friend of mine and I have been forming a band, and have been working together on pieces, a mix of classics and our own originals. We'll see where that goes...

Biology

My first real passion... I guess spending lots of time outside collecting bugs was a major factor for this. The other bias was growing up in a house with two doctors.

Some of the greatest contributors to modern biology, in my opinion, include Aristotle (eidos and genetic programs), Charles Darwin (on the Evolution of the Species), D'Arcy Thompson (On Growth and From), and Ernst Mayr (modern evolutionary synthesis). I also believe they do not do enough to teach these powerful, deep principles, something I experienced when I worked at BBN in educational technologies.

Ernst Mayr, the father of modern biology passed away on Feb 2005. I had the pleasure of meeting him and will always remember a discussion with him one beautiful June day on where biology was heading. His brilliance was that he was able to comprehend the full picture of biology from molecules to ecosystems, and cellular processes to evolutionary trends. His teleonomics is a key feature that separates living things from inanimate matter.

Mathematics (and Physics)

One of the most fascinating perspectives I have ever read in Mathematics is the relation between geometric structures and dynamics systems. I've listed a few books and references

• The Road to Reality , R. Penrose, Wikipedia summary. An excellent read (give yourself 12 months) that ties geometry to quantum physics and a bit of String theory is: (some good reviews) The background image on this page is of a "Hopf Fibration", a bundle of S1 fibers that pack perfectly with no intersections in a S3 space (surface of spheres are S2). Penrose's 'Twistor' model has this configuration for all the coincident rays of a twistor.
• Nature of Space and Time - This is a classic dialogue between Roger Penrose and Stephen Hawking.
• The Mind of God - One of Paul Davies' fascinating books that looks at how math, physics, biology, and computing are all interdependent on each other. I personally feel these kinds of views will achieve more understanding for where we come from than any current religion can.
• Wheeler'sDelayed Choice Experiment-- for some nice mind-bending thoughts.
• The famous EPR (Einstein, Podolsky, and Rosen) Experiment-- disorts our view of causality, but doesn't break it.

Art

I enjoy most forms of art, but I have a strong fondness for Impressionist Art. A favorite of mine is Monet's beautiful Giverny series, which portends the transition from impressionism to abstract art. At the core, it's about color and sentiment, and the association of feelings to images and locations.

Whenever I can find a few days of uninterrupted time, I'll work on sketches and drawings using the following media: Pastels, Pencil, Watercolors. I suspect this is something I'll be pursuing again more later in life. Here are a few I have done over the years.

History and Civil War Interest

I am a great fan of Thomas Jefferson (we share the same birthday), and always appreciated his broad interests on so many topics. Wish there were more people like him today in D.C.

When I read the Killer Angels by Michael Shaara, I became fascinated with the Civil War and the people who fought in it. After Michael's death, his son Jeffrey continued his work and wrote both a prequel (Gods and Generals) and a sequel (The Last Full Measure) to his father's novel, creating the trilogy.

I have a deep sense that in both civilization building and national conflicts humanity somehow ends up re-enacting the unthinkable. Even our current war in Iraq follows the same path of follies as preceding generations-- we're only fooling ourselves if we think we can have the outcomes we want by not heeding history.

Publications

• The Semantic Web in Action, Feigenbaum L, Herman I, Hongsermeier T, Neumann E, Stephen S, Scientific American December 2007.
• Semantic Web-based Data Representation and Reasoning Applied to Disease Mechanism and Pharmacology, Qu AX, Gudivada CR, Jegga AG, Neumann EK, Aronow BJ, 2nd IEEE International Data mining and Bioinformatics 2007 513.
• Gudivada CR, Qu AX, Chen J, Jegga AG, Neumann EK, Aronow BJ, Identifying Disease-Causal Genes using Semantic Web-based Integration of Genomic-Phenomic Data. Journal of Biomedical Informatics Journal 2007 (Accepted).
• A Genome - Phenome Integrated Approach for Mining Disease-Causal Genes using Semantic Web, Ranga Chandra Gudivada, Xiaoyan A. Qu , Anil G Jegga, Eric K. Neumann , Bruce J Aronow, HCLS Workshop, Banff WWW2007
• Knowledge networks in the age of the Semantic Web, Eric Neumann and Larry Prusak, Briefings in Bioinfomatics, May 2007
• Advancing translational research with the Semantic Web, A Ruttenberg, T Clark, W Bug, M Samwald, O Bodenreider, H Chen, D Doherty, K Forsberg, Y Gao, V Kashyap, J Kinoshita, J Luciano, M S Marshall, C Ogbuji, J Rees, S Stephens, G T Wong, E Wu, D Zaccagnini, T Hongsermeier, E Neumann, I Herman and K-H Cheung, BMC Bioinformatics 2007, 8(Suppl 3):S2
• The Advantages of a Drug Safety Commons, Bio-ITWorld, Oct 2006
• Freeing Data, Keeping Structure, Bio-ITWorld, June 2006
• RDF- The Missing Link, BioITWorld March 2006.
• Biodash: A Semantic Web Dashboard for Drug Development in Pacific Symposium on Biocomputing 2006
• Finding the Critical Path, DDW-Online 2005
• A Life Science Semantic Web: Are We There Yet? , STKE, 10 May 2005
• Davidov E, Clish CB, Oresic M, Meys M, Stochaj W, Snell P, Lavine G, Londo TR, Adourian A, Zhang X, Johnston M, Morel N, Marple EW, Plasterer TN, Neumann E, Verheij E, Vogels JT, Havekes LM, van der Greef J, Naylor S. Methods for the differential integrative omic analysis of plasma from a transgenic disease animal model. OMICS. 2004 Winter;8(4):267-88.
• Eric Neumann, Eric Miller, John Wilbanks, What the Semantic Web could do for the Life Sciences, Drug Discovery Today: BioSilico 2(6): 228-236 (2004).
• Eric Neumann and Jeffrey Thomas, Knowledge Assembly for the Life Sciences, Drug Discovery Today vol. 7, No. 20: S160 Ð S163 (2002).

Semantic Web

I got caught up in semantics while a researcher at BBN. My background in scientific research made me very cautious how scientific information was being represented in computers. In many cases, people well-versed in computer engineering and data representation were constantly struggling with how domain-experts related to information. Often a gap would arise between what was in a scientists mind and what ended up on the computer. That's when I started reading about information semantics and what it could offer to intelligent systems.

While working on an NSF funded project called 'GenScope*' to model biological phenomena across multiple levels of scale, we began trying to encode the various biological structures using an extensible language called 'Genetica' and storing it as an early form of XML that was tagged within an HTML page, in order to embed the model information in web pages. Around 1994 I serendipitously met Tim Berners-Lee in my home town and began pestering him on what were valid ways of representing biological knowledge using semantics. At about this time he was also composing the Semantic Web, but when he described it to me I could not make neither heads nor tails of it back then.

As I became more involved in bioinformatics and the various attempts to merge and analyze data, the concept of ontologies became more and more prevalent. In 1998 at the Montreal ISMB at a Birds-of-a-Feathers meeting, David Benton (Smith-Kline) and I proposed starting a Bio-Ontologies consortium to try and address the various hot topics. It was at this meeting that Michael Ashburner had openly proposed his Gene Ontology initiative, and told the audience he "had no time for arguing about concept formalisms on issues such as Thing or God". Subsequently, his version of a Gene taxonomy took-off with the genomics community, while more formal groups focused on headier issues.

I was at the time completely pre-occupied with how to stitch molecular and phenotypic data together to uncover patterns that might explain mechanisms of normal function or diseases. It wasn't merely to organize your data in larger and larger tables, using complex schemas-- that was the problem I noticed back at BBN, where the scientist was left out of the formalism. At about this point, the ideas TimBL had told me back a few years earlier started coming together. It especially made sense in the context of trying to put together molecular pathway models, which are really semantically-defined graphs. In 2000 we took a stab at representing these using the N3 short-hand, and loading them into his CWM RDF-tool (see Examples). From this point on, I was a Semantic Web convert.

Along with Vincent Schachter Aviv Regev, and Joanne Luciano, we began the Biopathways Consortium, and as I saw it, the modeling of complex systems required the same tools as used in semantic modeling. When the BioPAX exchange language was proposed at ISMB 2002, I made sure that OWL would be at the center of it because of its major advantage over XML in representing graph and logic information. At the same time, I was looking for a data-graph storage system for my company's (BG) informatics needs. We had originally worked with the Incellico technology, but needed to look for something more scalable.

TimBL was also not too happy with the lack of commercial offerings for data-graph storage. At the same time, Susie Stephens, then at Oracle, tried to get BG to use a new graph technology called NDM-- we agreed to test it, and within 6 weeks we had our systems biology data uploaded into it and working great! I then made the off-the-cuff suggestion to Susie of contacting TimBL who was looking for a graph storage database and might find NDM quite interesting. After a few calls between TimBL, Eric Prud'hommeaux, Susie and the NDM tech team, a few POC evaluations were convincingly done, and eventually NDM transformed into Oracle's RDF solution.

Oracle's entry into the Semantic Web space has been a major reason why awareness of Semantic Web has begun to spread. Noteworthy here is that the connection made in trying to solve a life sciences data problem may have helped to jump start the recent Semantic Web movement. With the original web, Tim was addressing the needs of the physics community, but the complexities of life science and biomedical data, such as molecular pathways and diseases, will almost certainly require the Semantic Web.

Related Consortia and Projects

• W3C Healthcare and Life Sciences Interest Group
• BioPathways Consortium
• BioPAX
• Bio-Ontologies
• Drug Safety and Efficacy
• Clinical Observations Interoperability

Ontologies

• Disease Cards Onotology, Cincinnati Children's Hospital Medical Center
• Protein Interactions
• Protein Interactions

* For amusement, check out the ScienceMan review of the educational version of this software...

Semantic Web Demos

All of these demos are based on publicly gathered data that has been converted into RDF, but in some cases versions exist with richer sets of associated information. If you would like to see a demo of these, please contact info@clinicalsemantics.com.

• Clinical Trials Demo

  This demo uses generated SAS-table data on patients and adverse events.

• MutDB Demo

  This example is based on a set of data taken from Sean Mooney's MutDB Mutation Database.

• Kinase Demo

  This kinase demo was suggested by Hanif Khalak at Novartis, and uses public data from various sources. The data is also woven together with EBI's CGP mutation DB.

• Cancer Genes Demo

  This data set is taken from EBI's Cancer Genome Project (CGP). I intend to stitch this together with more cancer data from other sites as well.

• Chembank data demo

  Rendering of similarity search results from Chembank. Images generated through calls to Chembank from within Exhibit.

• View of HCLS Neuro Knowledgebase

  This demo was part of theHCLS demo shown at WWW2007 in Banff. I call NCBI's gene structure image generators from within Exhibit.

• PubChem data

  Set of PubChem data for agents targeting Parkinson's and Alzheimer's. From within Exhibit, I call the compound image generators at NCBI.