Mind to Market

Personalized Medicine at Medco

noreply@blogger.com (Steve Connolly) — Thu, 21 Aug 2008 23:47:00 +0000

In a recent press release, Medco Health Solutions announced that it is partnering with the FDA to study the applicability of personalized medicine methodologies to prescription drug treatment. Medco is the nation’s leading pharmacy benefit manager and the announcement signals the continued mainstreaming of pharmacogenomics in healthcare.

Medco and the FDA will jointly develop research projects, programs and strategies in the area of pharmacogenomics, the science of using a patient’s genotype to predict their response to drug treatments.

Medco’s objective is to advance pharmacy care and take some of the guesswork out of the prescription process. Specifically, the objectives of the program are to address:

Safety of prescription drugs
Physician participation in pharmacogenomic testing
Usefulness of tests in prescribing
Quantifying prescription information about drugs

Central to the research studies will be Medco’s extensive database of pharmacy claims. Medco is already involved with pharmacogenomic studies including Warfarin at the Mayo Clinic and tamoxifen with LabCorp.

Adam’s Next Big Venture

noreply@blogger.com (Steve Connolly) — Thu, 26 Jun 2008 05:25:00 +0000

Kudos to Adam Rubenstein, blogger extraordinaire, for landing a plum job at cutting edge pharmacogenomic firm ARCA biopharma. In addition to typing his fingers to the bone on seventeen blogs, Adam has been raising money for Fitzsimons BioBusiness Partners, getting MycoLogics off the ground and consulting with High Country Venture. Working at one place may seem like a vacation.

Good luck Adam!

The Progressive Blockbuster

noreply@blogger.com (Steve Connolly) — Thu, 22 May 2008 21:46:00 +0000

Retiring CEO of GlaxoSmithKline Jean-Pierre Garnier has published his thoughts on the shifting pharmaceutical business model in this month's Harvard Business Review. True to his roots, Garnier advocates returning the management of R&D to the scientists.

But Garnier doesn't stop at restoring R&D management to scientists; he also has some strategies for reorganizing the drug development process. The first is to separate the processes of creating first-in-class from best-in-class drugs. Although discovering a first-in-class may be more satisfying scientifically, pharmas incur less risk and equal, or greater, rewards from best-in-class.

These two R&D processes are currently intertwined in most pharmas which, according to Garnier, leads to reduced productivity in both processes. Of course many of the functions of the processes are shared; complete separation would be impossible, but improved performance could be achieved with a concerted effort.

Garnier's next idea is what he calls "progressive blockbuster." Blockbuster status is the goal that every pharma is seeking yet is becoming more elusive as we enter the smaller patient populations of the personalized medicine era. Enter the progressive blockbuster. This is a drug that has been tested on a carefully selected sub-group of patients which have been shown to benefit and have low side-effects from the drug. The FDA would approve the drug only for patients in this specific sub-group. Once the drug has been approved and marketed to this sub-group, additional sub-groups can be added one by one as they show good results and safety. Eventually you will be marketing to enough sub-groups to make it a blockbuster. Voilà: the progressive blockbuster.

It does sound a bit more palatable to the general population than the current situation; where you begin with a blockbuster population and wait for adverse events to occur. But correlating patient response to genetic profile or phenotype will remain a tough nut to crack.

The Promises of Web 3.0

noreply@blogger.com (Steve Connolly) — Thu, 15 May 2008 00:08:00 +0000

Put off by the unfulfilled hype of Web 2.0, some in the knowledge management community are now clamoring for Web 3.0. Dr. David J. Roberts, Chief Scientist at iBASEt, writes in Oracle's Profit Online that Web 2.0 technologies aren't really worth the bother and that the real value sought by enterprises lies in Web 3.0 technologies.

The compelling promise offered by Web 3.0 technologies is the ability to make inferences between contextually linked information thereby pulling new, creative combinations out of knowledge bases automatically. This is the holy grail of knowledge managers; to get machines to be able to reason, even just slightly, would offer a great deal of value.

With such a compelling value proposition as reasoning will Web 3.0 technologies render Web 2.0 worthless? There are still some very large obstacles to Web 3.0 as Roberts has described, i.e. machines can't handle ambiguity and major pieces of the Web 3.0 language (ontologies) have yet to be produced.

But is the interim value of Web 2.0 technologies really that low? Web 2.0 technologies provide integration for well defined processes, and it’s the fact that they must be well defined that renders them inflexible and in need of constant maintenance. Yet this integration is indeed quite valuable, too valuable to be left on the shelf until Web 3.0 is ready for prime time. The future world of Web 3.0 is indeed rosy but don't count Web 2.0 out just yet.

Consumer Genome Tests

noreply@blogger.com (Steve Connolly) — Fri, 09 May 2008 15:27:00 +0000

With the U.S. Senate passing the Genetic Information Nondiscrimination Act (GINA) last week the bill now moves on to President Bush for signature into law. One more impediment to the genomics age will be removed. Does this pave the way for personal genomics companies such as 23andme and Navigenics?

The Wall Street Journal took up the issue of accuracy with these consumer products last week. As with many new (although I can't really call it disruptive) technologies, this one is somewhat of an expensive novelty. The main criticism seems to be a lack of accuracy in identifying just what genetic diseases an individual may be susceptible to. Although there are numerous tests for susceptibility to heart disease, the current genetic tests don't provide much more accuracy than can be provided by measuring blood pressure.

Some fairly rare genetic diseases do show up clearly on the tests, e.g. a condition associated with a higher instance of blood clotting: factor V Leiden. There is a gray area where the tests will indicate a higher susceptibility to a disease but don't go so far as to say you will contract it. This would be valuable to an individual by allowing them to increase their level of monitoring and taking preventative measures.

So far the insurance companies have yet to buy in on the whole genome tests although they have been covering disease specific tests such as those for breast cancer prognosis.

What the personal genomics companies are selling now is information, information in a quantity and form with which scientists and clinicians, let alone consumers, can barely get their arms around. As Howard McLeod, professor of pharmacology at the University of North Carolina, is quoted as saying "You get a lot of information but very little knowledge." The real challenge, and true value, will be to extract knowledge from these tests.

Genetic Sleuthing

noreply@blogger.com (Steve Connolly) — Tue, 29 Apr 2008 04:48:00 +0000

No sooner did the U.S. Senate pass a bill barring genetic discrimination than the state of California announces that they will track down suspected criminals through the use of their relatives' genetic information. Certainly using genetic information to restrict access to healthcare and employment could create ambivalence towards the collection and recording of this information and the Genetic Information Nondiscrimination Act of 2007 (GINA) has gone a long way to reduce that. But now knowing that your genetic information can be used to find and convict a family member of a crime, with or without your cooperation, introduces yet another impediment to the collection of this information.

California is the first state to aggressively employ the new familial, or "partial match" search technology although several other states use it to a limited degree. So far California is only using DNA profiles within its own offender database; DNA taken only from convicted offenders, but plans are in place to extend the database to include arrestees.

The technology of genetic genealogy has progressed rapidly providing insight into human migration and ancestry. Apparently now the technology has progressed to the point where its use can be vital in a criminal investigation.

Although currently used to make exact matches between a DNA sample and a DNA profile in the database, the new policy allows the crime lab to tell investigators who might be related to the person who left the DNA sample. It's not enough for a conviction, but enough to identify, contact and sample the DNA of relatives of the suspect with the eventual goal of finding the suspect.

Familial searches have been common in the UK for the past few years due in part because of the over 4 million profiles of charged criminals and suspects. The practice is not without controversy and even Alec Jeffreys, the discoverer of DNA profiling, believes that using DNA from people who have never been involved with the law raises "potentially rather thorny" civil liberties issues.

Mining the Text Mountain

noreply@blogger.com (Steve Connolly) — Thu, 17 Apr 2008 05:19:00 +0000

I've had my doubts about text mining in the past; such as why we spend so much energy wrapping facts into coherent sentences only to have machines struggle to unwrap them. Not that there is any doubt that there is a huge amount of unmined text out there; decades of work for text miners. It's just that it seems of questionable value when this knowledge could be tagged as it were entered thus obviating the need to mine it later. But there is this vision, a vision that with sophisticated text mining tools some bigger picture can evolve from a mass of unstructured data.

In a recent article (subscription required) in Genome Technology William Hayes, director of informatics at Biogen Idec, alludes to this vision stating that writers cannot pre-tag the information while writing as the information can only be viewed within some larger context. The jury is still out on this but regardless of how knowledge is gathered and transferred in the future, we’ve got years of literature to mine from.

Hayes also claims that "almost all of our [biomedical] knowledge…is captured in the literature." I'm sure we can claim that researchers do publish their "good" results in the literature, but this is far less than our total knowledge. Because the process of writing and publishing knowledge is so difficult and expensive, only the very most promising results are usually published, leaving the vast remainder in lab notebooks, hard drives and researchers’ heads where it is difficult to distribute.

Larry Hunter, director of the Computational Bioscience Program at the University of Colorado, sees text mining as a way to improve current databases but remains skeptical that such systems can automatically keep a scientist up to date with the literature. Don't throw those reading glasses away just yet.

Right Brain, Left Brain Thinking

noreply@blogger.com (Steve Connolly) — Mon, 14 Apr 2008 18:11:00 +0000

For those of you who may be interested in an anatomical explanation of linear thinking there is an excellent video on the TED Web site of a recent talk given by neuroanatomist Jill Bolte Taylor, Ph.D. Not only has Dr. Taylor studied the brain in great detail but she also had the (mis) fortune to observe her own stroke first hand.

The brain, as Dr. Taylor shows us, is composed of two almost completely separated hemispheres. The analogy she makes is that the right hemisphere functions as a parallel processor and the left a serial processor. The right hemisphere operates in the present moment, thinks in pictures and processes sensory inputs.

The left hemisphere, the serial processor, is the half responsible for linear thinking. This part of the brain picks out details from the information streaming into the brain and categorizes and organizes it. It separates time into past and future. It thinks in language and is responsible for our ego.

Dr. Taylor's stroke occurred in the left hemisphere of her brain. Gradually, as the stroke progressed, she lost more and more of her left brain functionality and existed solely in a right brain world, a world where her language skills were lost, where her ability to carry out simple functions like dialing a phone were seriously impaired but a world where as her ego melted away she felt more connected to the world around her.

Personalized Medicine to the Rescue

noreply@blogger.com (Steve Connolly) — Wed, 26 Mar 2008 03:14:00 +0000

The story of pharmacogenomics has reached the mass media, a.k.a. the Wall Street Journal, as a means of stemming the tide of failed drug candidates beleaguering the pharma industry. The article cites drugs from Novartis AG, NitroMed, and Clinical Data that have saved drugs that had, or would have, failed to make it to market by targeting patient sub-groups based on genotype.

Of significance in the article is stating that the pharmaceutical blockbuster model is dead and that personalized medicine may represent the new paradigm. There are in fact upsides to the smaller markets that personalized medicines command, i.e. high switching costs. With a "one-size-fits-all" drug a patient can move from one medication to another with just a trip to the pharmacy, but when using a drug is customized to a patient’s genotype, switching to a drug that is known to be less effective will be a non-starter.

One local Colorado play that illustrates the resurrection of a "failed" drug is the in-licensing of bucindolol by ARCA Discovery. Although an earlier Phase III clinical trial of the drug was terminated due to lack of efficacy, evidence has shown that certain patient sub-groups that exhibit specific genetic polymorphisms can benefit from the use of the drug.

Although the NIH Roadmap Initiatives that promoted personalized medicine and biomarker approaches four years ago weren’t mentioned in the article, the cause has been taken up by the pharmas as a new way to fill those emptying drug pipelines.

Killer Semantic Apps

noreply@blogger.com (Steve Connolly) — Mon, 24 Mar 2008 15:43:00 +0000

Is coming up with a definitive application for demonstrating the utility of semantic analysis really that difficult? TextWise, a software development company in Rochester, New York, apparently thinks a good idea for this technology is worth at least the $1 million they are offering the winner of their SemanticHacker $1M Innovators Challenge.

The rules of this contest require the contestant to use, or propose to use, the SemanticHacker API, based on TextWise's Semantic Signatures^® technology, to develop a software application for a specific industry vertical. Although it is up to the contestant to propose a vertical, TextWise suggests industries such as "healthcare or pharmaceuticals might be good places to start." Wonder who tipped them off?

As explained in TechCrunch, Semantic Signatures^® uses natural-language processing to extract relevant terms from text then applies semantic analysis to automatically categorize Web pages. Not a bad idea, but the technology can be a bit flaky. Semantic Signatures^® used Wikipedia as a reference; connecting the concepts extracted from the text and matching them to Wikipedia articles.

One of the cornerstones of the W3C specification for the Semantic Web is its use of Web Ontology Language (OWL), although OWL only specifies the format of the ontologies there is the assumption that human domain experts will be required to accurately develop an ontology. TextWise claims that ontologies developed in this way "do not align with customer needs and…rapidly become obsolete." Perhaps, but without some agreement on the ontology all you have is a folksonomy, which reduces its value in collaborative efforts.

The Holy Grail that drives the concept of semantic analysis is the ability for the software to do "connecting the dots" process that is normally done by humans. We humans can juggle a few thousand "dot" in our heads but connecting one to another, or maybe some complex combination of five to another twelve, gives most of us a headache. And when you start thinking about connecting a million or more dots, well, time to start thinking about a simpler project, like brain surgery.

Helicos Looks to Collaborate

noreply@blogger.com (Steve Connolly) — Fri, 14 Mar 2008 16:39:00 +0000

No sooner did Helicos BioSciences announce the shipment of their first Single Molecule DNA Sequencer did they announce the formation of a federation of IT vendors to handle the crush of data that the new machine will be churning out. In an interview in the online edition of Bio-World, Kristen Stoops, Director of Informatics at Helicos, describes the astronomical amount of data that is collected by their machine; upwards of 5 TB per day. A next-gen sequencer obviously requires next-gen data management.

They are looking for hardware, networking technology, high-performance computing, systems integrators, bioinformaticians, basic IT infrastructure companies, and LIMS vendors. Then get this diverse group to integrate their technologies for as Stoops says "speed…is going to be key in mitigating the data management challenges."

Helicos' eventual goal is clear: the $1000 whole genome, but to get there they’ll have to speed up their process from 1 billion bases per day to 1 billion bases per hour. Users of next-gen sequencers are cautioned to have a clear idea of the data management issue, this will in fact be a major hurdle in the adoption of these sequencers.

Shifting Pharma Business Model

noreply@blogger.com (Steve Connolly) — Tue, 26 Feb 2008 17:19:00 +0000

Marc-André Gagnon's recent article in Genetic Engineering News is an observation of a trend in the big Pharma business model, a trend that he feels particularly disappointed with. That trend is one in which big Pharma does less drug research and development and more marketing. Although Gagnon is a bit peeved with this, it is nevertheless inevitable and irreversible.

In a talk two years ago, Steven Burrill went so far as to predict that the big pharma R&D labs of not too distant future would be empty while the marketing department would be booming. The economic reality of rising risks and corresponding costs of developing new drugs along with the shrinking size of markets due to segmentation of diseases has left big pharma without a clear future. One thing is clear: pouring ever increasing amounts into internal R&D is not yielding the returns they used to.

A new long term business model may indeed be one in which smaller, more agile biotechs perform the early stage drug development, out-licensing to larger pharmas once success has been shown. Large pharmas could then leverage their brands, sales and marketing and distribution systems to commercialize these drugs. This is in fact what Gagnon is observing; R&D budgets are shrinking relative to sales and marketing.

The driver behind this trend is the investment community; what business model can big pharma adopt to continue the returns they have achieved in the past? The Dow Jones US Pharmaceutical and Biotech Index continues to drop as the industry searches for a model. Extending the life and markets of existing drugs through promotion and minor modification may offer an interim solution until a more sustainable solution is reached.

But assuming that big pharma will continue to innovate scientifically simply because they have been successful with it in the past or that it has been a compelling marketing message is steering toward an economic downfall. The innovation must be within the business strategy; to capitalize on the current and future strengths of the industry.

Early Stage Funding in Israel

noreply@blogger.com (Steve Connolly) — Mon, 18 Feb 2008 02:11:00 +0000

I had coffee recently with Amir Genosar, a very talented and creative Boulder entrepreneur. Amir has no less than three medical device start-ups underway; including aespironics and SteadyMed.

Having started his career in Israel, Amir now calls Boulder home. Why then does he, while surrounded by Boulder's angel and VC investors, still rely on Israeli investors to fund his businesses? Amir contends that Israeli investors are willing to accept more risk in return for lower valuations. These investors are more interested in the technology, even if unproven, than they are in the business model.

A business plan submitted to an Israeli investor for early stage funding must be heavy on technology and light on marketing and business model. How could this be? Are Israeli investors that unsophisticated? I have a few theories: there are many sources of funding for product development in this country such as the federal SBIR and STTR programs, state funding programs to commercialize university IP, corporate R&D or the entrepreneur's own resources. If a product does not receive funding from one or more of these sources, how viable could it be?

There are certain categories of products that provide exceptions to this. In new and evolving markets such as Web 2.0/social networking where products traverse their entire life cycle in a period of weeks, investors may take on less developed products simply to get in on the action. No federal program could possibly keep up with the pace of development in this arena, but the impressive valuations of Web 2.0 companies are hard for VCs to ignore.

In categories such as medical devices and enterprise software however, the number of developed products is high providing ample opportunities for investors to forego the risk of product development. Although valuations are higher, more capital is available than in a small country such as Israel.

It wouldn't surprise me that Israel does not have government programs that support commercialization and that this is a role taken on by private investors. Given the non-dilutive product development resources available to the entrepreneur in this country it may be less desirable for the entrepreneur to go the Israeli route.

Genomic Health’s Healthy Results

noreply@blogger.com (Steve Connolly) — Fri, 08 Feb 2008 06:09:00 +0000

Genomic Health's year end revenues for 2007 had more than doubled over 2006 it reported last Tuesday. A good 98% of that revenue was from their Oncotype DX multi-gene assay that predicts the recurrence of breast cancer in newly diagnosed patients. Both payers and physicians appear to be adopting the technology; an early entrant in the personalized medicine arena.

Genomic Health claims that health plans covering greater than 70% of the population now reimburse for Oncotype DX. Since the test provides information on which patients will benefit from chemotherapy, payers can reduce the costs of unnecessary chemotherapy treatments through the use of Oncotype DX.

Regardless of the economic benefits of the test, patients that will not benefit from chemotherapy are spared the discomfort of that treatment. Personalized medicine technologies appear to have quantifiable benefits but that alone is not sufficient to predict how well they will do in the marketplace. With these numbers however, Genomic Health sounds more confident and predicts revenues to increase to between $100 and $110 million for 2008.

Translational Research in Colorado

noreply@blogger.com (Steve Connolly) — Fri, 25 Jan 2008 23:38:00 +0000

The Colorado BioScience Association kicked off the first BioBreakfast of the year at the new Education 1 building on the Anschutz Medical Campus featuring University of Colorado's new Vice Chancellor for Research, Dr. Richard Traystman. Dr. Traystman together with Dr. Ronald Sokol of The Children's Hospital, presented CU's bid for a Clinical and Translational Science Award (CTSA) grant.

The CTSA grant is part of the National Institutes of Health's Roadmap for Medical Research, a program spearheaded by NIH Director Dr. Elias Zerhouni in response to the rapidly depleting drug pipelines of the nation's pharmaceutical companies. The "T" in CTSA stands for Translational; the concept of bringing a drug from the laboratory bench to the patient's bedside. In other words: applying biomedical research to finding new therapies.

Now we all assume that biomedical research is a good thing and that it's where all of our new therapies will come from, but the reality is that even as the NIH's budget doubled from 1993 to 2004 the pharmaceutical pipelines were thinning. We found out that there is no correlation between spending on basic research and new treatments being brought to market.

The CTSA program represents one of the ways the NIH is seeking to address this issue by becoming more proactive in generating therapeutic results instead of journal articles. The General Clinical Research Center grants are being phased out in favor of the new program, compelling current GCRC recipients, such as CU, to retool and apply for the new program.

Final decisions on the application will be received in May. We’ll keep our fingers crossed.

Investing in Theranostics

noreply@blogger.com (Steve Connolly) — Tue, 22 Jan 2008 17:25:00 +0000

Making a strong case for the commercial development of personalized medicines, Lisa Haile, JD, PhD calls for greater investment on the part of venture capitalists and pharmaceuticals in a recent article in Genetic Engineering News. Theranostics is the term coined for combining diagnostics and therapeutics; an essential part of delivering on the personalized medicine promise.

Not only will theranostics reduce the risks of developing drugs, Dr. Haile states, but lower the costs of development as well. Yet as compelling as that sounds, she writes that the number of commercialized theranostics products today is unacceptably low and lays the blame at the feet of the VCs and pharmas.

Personalized medicine is no doubt a paradigm shift from the "old" blockbuster pharma model, a model that pharma is loathe to abandon. More than a paradigm shift, this may in fact be a technology disruptive to the industry. As in the case of many disruptive technologies, they do not fit nicely into the industry's business models and as a result are usually passed over for technologies that provide the promise of blockbuster status.

I don't think all VCs are oblivious to the promise of theranostics however. Steven Burrill introduced me to the term two years ago in a talk where he was singing its praises. Atlas Venture, Boulder Ventures, Skyline Ventures and InterWest Partners have backed a start up pharmacogenomic play called Arca Discovery to develop cardiac theranostics. As is often the case with a disruptive technology, it requires an entrepreneurial effort to make the initial forays into commercialization.

The Curse of Knowledge

noreply@blogger.com (Steve Connolly) — Mon, 07 Jan 2008 04:46:00 +0000

A recent article in the New York Times entitled "Innovative Minds Don’t Think Alike" by Janet Rae-Dupree really misses its mark. Instead of innovation, what Ms. Rae-Dupree seems to be referring to is market acceptance which is another subject entirely. Granted, the commercial world really has little value for innovation for innovation's sake, but describing a product as not being innovative because engineers can't describe it properly to consumers is just missing the whole concept of innovation.

Now don't get me wrong, I'm not a big fan of cool but useless technology, but an innovative idea is not necessarily a useful one. That being said, often times a "useless" technology will morph into something more useful and even commercial over time. Examples: too numerous to list.

Both collaboration between stakeholders and feed-back from the market are indispensable aids in guiding product development. But they can stop innovation dead in its tracks. The constant need for buy in from stakeholders and the market while the idea is developing will stymie its development and can reduce it to a merely incremental one.

An ever present danger in developing innovative technology solutions is that you become so entrenched with the solution that you can no longer talk about it in layman's terms. You think you are describing it in terms an eighth grader can understand when in fact most Ph.D.'s in the field think you're speaking in a language somewhere between ancient Etruscan and Klingon.

But this is actually a good thing as far as being innovative. If everyone understood your "innovative" product upon hearing about it for the first time how innovative could it really be? Either you've got your message down perfectly or it's just too obvious.

But Ms. Rae-Dupree's premise is that if laymen can't understand the product, even in the development stage that it is anti-innovative and you are definitely off track. Again, maybe a quick check with a definition would help. Real innovation happens in committees? If this were true the U.S. government would be the most innovative organization around.

Certainly one impediment to non-linear thinking is that you'll be met with a lot of skepticism and blank looks. Take this as a compliment; disruptive ideas are by their very nature unobvious and people are naturally resistant to ideas they don't understand. In fact they can be down right hostile.

There is no doubt that market acceptance is the goal for a commercial product whether innovative or not. But if you are developing a truly innovative product, a disruptive product that will introduce a new paradigm to the market, then don't expect everyone to understand it. The trick is to maintain just enough connection to the market to be real without corrupting the creative process.

Dark Matter of Science

noreply@blogger.com (Steve Connolly) — Thu, 03 Jan 2008 16:42:00 +0000

A scientist sets out to show that a gene is linked to a disease and runs an experiment to prove it. But things don't work out as planned and results show that the gene has no connection to the disease. Scrap the results and head back to the lab? This is in fact what most researchers do; putting aside "negative" results in favor of results from experiments that show "positive" results. But are negative results any less valuable than positive ones?

Although what tends to be exciting in science are the conditions in which one entity affects another, the lack of affect can be useful in pointing a researcher away from an invalid model and toward a more accurate one. But if these negative results are not published or transferred to the scientific community there is nothing explicit to prevent other researchers from going down the same dead-end.

With the costs of publishing, both in labor and capital, can publishing negative results be justified? Many journals do not publish negative results per se, although many articles published with positive results may cite negative results that were encountered along the way. A handful of journals have recently been established for the sole purpose of publishing negative results. The Journal of Negative Results in Biomedicine, the Journal of Negative Observations in Genetic Oncology (NOGO), and the Journal of Negative Results Ecology & Evolutionary Biology are three in Biology. It's interesting to note that NOGO puts the cost of an average negative study at between $5,000 and $20,000 per gene investigated. This certainly provides economic incentive for publishing negative results.

As science moves into the informatics age where knowledge is more readily accessible and distributed, value differences between "positive" and "negative" results will undoubtedly diminish. Although no one will ever describe a negative result as a "breakthrough" no breakthrough has ever been achieved without its preceding dead-ends.

Modeling Biology

noreply@blogger.com (Steve Connolly) — Tue, 18 Dec 2007 05:30:00 +0000

As someone who promotes the ideas and technologies of computational biology, I am regularly confronted by skeptics who either doubt the value or the capabilities of computational models of biology. There is a perception of a computer either churning for days to determine what a trained scientist can figure out in a few minutes or the inability of the program to provide a correct result to a seemingly simple problem.

Perhaps this is an issue of setting expectations. The cases above expect either too much or too little from computational modeling. As a first pass a computer model should be expected to perform those tasks that are already being done manually by scientists. Every scientist, even those who have never written a line of code, has some mental model of the biological systems she is working with. The mental models may be complex and subtle, i.e. difficult to express explicitly, but they can be expressed nevertheless.

Terry Quatrani writes in a recent Dr. Dobb's Journal article entitled Agile Modeling: No, It's Not an Oxymoron, that computer models are useful both by enabling communication with other team members and by focusing the thought process. She points out that 93 percent of agile team members sketch on the whiteboard, definite proof of modeling. If you're a scientist sketching on a white board or notepad, you're a modeler.

So why bother with the arduous process of converting those sketches into millions of lines of code? Although the sketches may contain genius, they are difficult to transfer, reproduce, extend, integrate or eventually, simulate. Although they contain the essence of thought, they lack the conveniences and efficiencies that modern information technologies provide us with.

My Family Health Portrait

noreply@blogger.com (Steve Connolly) — Sat, 08 Dec 2007 18:30:00 +0000

Just in time for the holidays, the Office of the U.S. Surgeon General together with the National Human Genome Research Institute (NHGRI) has developed a Web site to organize and display family health history: https://familyhistory.hhs.gov/. Despite the overwhelming numbers of Americans who believe that knowing their family's history is important to their own health, very few have collected and organized this information. This new software makes it easy to input and report this information.

No information is stored on the site; the user downloads it at the end of the session. You may even download the software for use on your own computer. Outputs are either in report form or in a graphical family tree structure. At this point you may wish to conduct your own genetic analysis or take the report to your family doctor which will provide her with greater insight into your predisposition to genetically linked diseases.

Marathon Dog

noreply@blogger.com (Steve Connolly) — Sat, 17 Nov 2007 00:14:00 +0000

A few years back I had the opportunity to take a friend's dogs out for runs. One of these dogs was a bulldog named Bandit. Bandit was a bundle of fast twitch muscle and would literally pull me for first mile. The second mile he would jog beside me contentedly, the third I would have to coax him along and by the fourth mile I would have to sling him over my shoulder if I had any chance of making it back home.

Despite this seeming lack of endurance, given enough time and doggy treats, I was confident I could train him for a marathon. Fortunately for Bandit, his owner decided that 5k's were the limit and probably spared him an early demise.

The story of my dog training came to mind at yesterday's Healthcare Heroes event sponsored by the Boulder County Business Report. A panel of healthcare experts provided their opinions on the future of healthcare and fielded questions. An audience member asked whether it was appropriate for healthcare providers and payers to incentivize healthy activities such as diet and exercise, since it would potentially lower the individual's healthcare needs. Although some payers do offer incentives to reduce harmful habits such as smoking, John Sackett, CEO of Avista Adventist Hospital, provided a warning that it would be difficult to issue these incentives across the general population. Given the wide genetic disparities in populations, Sackett said, standard incentives such as weight loss, may not be helpful and may even cause harm.

One size fits all healthcare has proven to be an inadequate model, should preventative programs prove any different? Are genetically based incentives a possibility? Navigenics has announced that it plans to launch a genetic test called Health Compass that will indicate what lifestyle changes people could make to avoid or delay disease. This test will be offered via the Internet for $2,500.

Although these types of tests are in their infancy, they may very well become a standard part of the healthcare process in years to come. Bandit didn’t need a $2,500 test to tell him to stop at 3 miles, his genotype was expressed quite clearly. With humans it's not nearly as clear and the benefits of testing my prove to be compelling.

Google's OpenSocial

noreply@blogger.com (Steve Connolly) — Wed, 14 Nov 2007 05:44:00 +0000

I've been investigating Google's new open standard for social networks: OpenSocial, is this a big step forward in integrating social network sites or an attempt to block Facebook's advance? There is no doubt that developers are faced with a confusing assortment of APIs to negotiate if they are going to link information from one Web site to another. And the information these APIs present is fairly standard stuff: names, demographics, relations, so why shouldn't this be standardized to make it easier to access?

Well for one thing Facebook is huge and growing bigger every day. They have an API which plenty of developers use. The big kid doesn't need a standard; it's the runners up that need one if they have any hope of staying in the game. And Google is certainly playing catch up here.

In many ways Google's initiative is similar to the caBIG initiative set up by the National Cancer Institute three years ago. Concerned about the proliferation of data silos among their grant recipients that hindered sharing and access to data, the NCI set up the Cancer Biomedical Informatics Grid (caBIG) as a way to standardize APIs among institutions. NCI has even set up a certification process to rate institutions on how well they comply with these standards.

An unfortunate downside to this standardization of APIs is the inability to access specialized functions or data that may reside at the data source. The standardization committee can be petitioned to extend the standard, but it is inevitable that not all terminology will be included. In some cases the data source will have their own API through which developers can access the non-standard information.

Social networks have a long way to go before they reach the level of complexity of cancer biology. So imposing a standard now may not be a big restriction on the information. At least developers will have now have a choice; use the OpenSocial API to access all social networks or site specific APIs if there are some specialized information to retrieve.

Web 2.0 for the Enterprise

noreply@blogger.com (Steve Connolly) — Mon, 05 Nov 2007 17:29:00 +0000

While social networking sites and wikis are rapidly gathering a large following the question of what value these technologies hold for business enterprises remains largely unanswered. Information Week has provided a useful summary of the adoption of various Web 2.0 technologies in the enterprise. In general they feel that much of the Web 2.0 functionality can be provided by existing technologies such as SharePoint and content management systems.

The highest value Web 2.0 functionality was that of collaborative content tools with 39% of responses from an online poll finding it to be important. Notably the social networking functionality of Web 2.0, the functionality the provides the value behind facebook and myspace, seems to be the least valuable feature to enterprises, only 5% of responses from an online poll found it to be important. But are they really distinct features? The real value of Web 2.0 is its connectivity; you can shift the focus of what you find valuable, whether it be practical solutions to business problems or who your favorite band is, the point is that it is networked and available.

The idea of focusing on those features that provide real business value is a good one, but focusing to the point of information silos is contradictory to Web 2.0. The ability to provide focus will be provided by better semantic integration, technologies that will allow a user to extract highly relevant information quickly rather than to sift through mounts of semantically mismatched information currently retrieved with non-specific searches.

Predicting the Future

noreply@blogger.com (Steve Connolly) — Tue, 30 Oct 2007 21:06:00 +0000

If we could accurately predict what the price of a publicly traded stock would be tomorrow we would stand to make millions. We all know that we can't do that or even simpler predictions. But there are more general predictions that we can make that are just obvious; drugs will continue to improve to provide better healthcare benefits, computers will become better integrated to provide a more seamless interface with our lives.

I've been reading Neil Stephenson's book Snow Crash; an action, adventure, video game like view of the future. Probably not one of Stephenson’s best, yet the thing that struck me was that it was first published in 1992, really before the modern Internet, before the advent of serious computer virus' and certainly before Second Life, yet it's all here as if he had some portal into the future. Granted many of his predictions have yet to fully come to pass, he was probably shooting for 2020, but some of them are well on their way.

One prediction that Stephenson and other futurists make without fail is the ability to access multiple data sources seamlessly, i.e. by using a single interface users can quickly and easily access a multitude of data sources and mix and match the gathered data. It seems so effortless and so obvious. Yet in the real world the notion of this level of integration would be met with questions regarding the substance in your cigarette. We have created so many isolated and discordant systems that the effort of tying them together seems hopeless.

This is why we can marvel at systems such as Google which appear to make sense of it all. These search engines can seemly extract the needle from the haystack, which is in fact amazing, but falls far short of the real value of the Web. Far too much manual munging is required to get the information into something even remotely integrated. Start your own database and integrate to it? A workable solution, but just a drop in the information ocean.

If the rest of the Web insists on remaining silos, at least some visionaries are creating integrated environments, i.e. facebook. facebook is creating an integrated Web within the Web and members seem quite content to stay within its bounds. Relatively small now, it will continue to grow rapidly as third party applications add more value to the integrated environment. An entertaining toy? Maybe, but it may provide a glimpse of how other domains may provide integration.

Moving Forward with Personalized Medicine

noreply@blogger.com (Steve Connolly) — Fri, 26 Oct 2007 03:48:00 +0000

A recent article in Genetic Engineering News highlights the progress Personalized Medicine is making in the healthcare environment and some obstacles that still must be overcome. Although the FDA and health insurance companies are beginning to open up to the possibilities of personalized medicine, this is far from being a common medical practice.

The one shining success in PM is the ability to distinguish between different types of breast cancer based on the expression of the HER2 gene. The results of this test indicate which therapy will provide the best treatment for the patient. One reason for this success has been an absolute convergence between diagnostic testing, devices, drugs and information.

The realization that the current model for drug development will not yield profits when applied to the smaller populations served by PM therapies is beginning to sink in. Yet no viable alternative business model for pharmas exists. There is the thought that PM is still in its infancy, that the blockbuster model still has a few more years left in it. Pharma is trying out some innovative new business models but no clear success has been achieved.

There are also regulatory hurdles to overcome. Very few FDA guidelines for clinical trials of PM therapies exist. PM is also going to require some very highly integrated information systems, combining the whole of the biomedical knowledge base with the patient's own electronic record.

Consensus is that PM will be the norm by 2020 if all stakeholders; regulators, policy makers, and private industry can work together on it. The benefits are compelling but the obstacles daunting.