Up at 5AM: The 5AM Solutions Blog

Imagine you've just been told you have a partially obstructed artery that requires surgical intervention. Your doctor tells you he could treat it with angioplasty, a procedure to widen the vessel, or coronary bypass, a procedure that, well, bypasses the blockage itself by grafting in additional vessels to route around it. Recognizing this will have life-changing consequences, you ask your doctor which is more effective. Here's the surprising thing: Until a few years ago, whatever his answer, it would have had more to do with his training, experience and intuition than it would to do with an objective comparison of the effectiveness of the two treatments. The reason that is so is explained in "The Best Medicine", a fascinating article by Sharon Begley in the latest issue of Scientific American.

As Begley lays out in her article, to compare the effectiveness of two or more treatments involves running expensive, multi-year, double-blind trials with patients randomly assigned to one treatment or the other. The cost to perform these controlled trials runs into the millions, often costing even a hundred million dollars or more. Given the costs of comparing treatments and the number of treatments needed to compare, the reality is that most decisions about which treatment to use in a given situation are left to the best judgment of the provider involved. Referring to the guidelines provided by the Infectious Disease Society, Begley notes that more than half of recommendations are based on expert opinion alone.

Well, so what, you might say. If it really costs so much to compare the treatments, in most cases, expert judgment is probably good enough, right? I mean, sure, if you're having heart surgery, you might want to know for sure, but what about treating a cold or high blood pressure? Isn't the doctor's judgment good enough?

So, here's where your pocketbook gets involved.

One expert quoted in the article (Elizabeth A. McGlynn from Kaiser Permanente's Center for Effectiveness & Safety Research) estimates that we're spending as much as a third of our health care dollars for treatments that are unnecessary or ineffective. This isn't small change we're talking about. A third of our health care spending amounts to close to $900 billion dollars a year. As Begley highlights in her article, with health care costs in the trillions and growing, one way to bring spending under control would be to pay only for the most effective treatments... If only the cost of finding out which treatments are the most effective weren't so high.

Here's where the good news begins. In a bright idea that began with David J. Magid, director of research for Kaiser Permanente's Colorado Permanente Medical Group, researchers are beginning to turn to the data locked in electronic health records to answer the same questions that controlled trials have answered--at a fraction of the cost. The practice is known as comparative effectiveness research, and it's already yielded impressive results. In the case of Magid, with only $200,000 he was able to determine in only a year and a half which drugs patients should use to supplement their blood pressure treatments if a diuretic alone wasn't sufficient. Still not chump change, but certainly a fraction of the $120 million spent on an earlier controlled study about the effectiveness of hypertension treatments, and a minute fraction of the estimated $3.1 billion dollars that were being spent on unnecessary medications as determined by the earlier controlled study.

The thing that stands out to me about the researchers described in Begley's article is that they are looking at their challenges through a different lens and finding new ways to make use of the resources they have available to them to answer questions that matter. Reading about them excites my imagination. What they are doing makes people's lives better. I wonder if there isn't some way to extend their ideas. What if the data in electronic health records were enriched with data from publicly available research databases? Would new insights about treatment emerge? What about opening up the data locked in institutional databases to hobbyists and outside researchers? Would collaboration and crowd-sourcing lead to more rapid research? I'm certain any of these ideas would entail technical, legal, and political challenges. But what would they enable? The impressive performance of comparative effectiveness research hints that the gains might be significant. It really feels like we may be on the edge of something big.

As you probably know, this week the FDA released examples of new warning labels for cigarette packs. The images are graphic – they include an awful photo of a scabbed mouth with rotten, discolored teeth; diseased lungs; smoke emerging from a tracheotomy hole; and a scarred and stitched corpse. They're just yucky.

Will they stop people from smoking, or encourage smokers to quit? I like to think so – not only selfishly because some people I love smoke, but because I believe in the power of visual communication.

Compare the two "images" below. The first is the current warning about cigarette safety. The second (obviously), is the corpse.



Quick – put your hand in front of those images so you can’t see them. Can you recall the words on the left? Can you recall the message on the right? Now I know there are plenty of people who can probably recite that standard written warning message from memory, seeing as it’s been around for 26 years. But seeing these side by side does make a point about how compelling images can be.

And I'm a reader, so I say that with some reserve. At museums, I'll laugh at myself for spending nearly equal time appreciating the art as reading the captions describing it. Words are extremely powerful, but images can express even complicated concepts quickly and cleanly. At 5AM, we develop software for complex biomedical challenges – and we'd never be able to write a lick of code (or maintain our record of 100% deployed-and-used software) without visual artifacts being essential to our process.

Modeling is the obvious case in point. Would you prefer to read about how subsystems interact, or how software is deployed, or view it in a UML model? Even if you don't know UML, the simplicity of the images make it pretty comprehensible. How can they not be, with stick-figure "actors" representing people, straightforward lines to indicate "some kind" of relationship or interaction between items, and boxes with labels to identify parts of the system (classes, interfaces, components – even if you can't tell the difference, you still get the concept). Modeling is at the core of our work, because not only does it simply and powerfully express (to our clients and our development team) "what happens," but because the very act of drawing a model forces us to think abstractly, and to design.


(I specifically show a simple example here so that those who don't know UML can go visit IBM/Rational's nice UML site.)

Fads and fashion in software development are constant, but whether you're "MDA all the way," or "scratch it on a whiteboard then crank the code," modeling should be a constant too. You're going to have to explain how your software works at some point, and "read the code" isn't the right answer (although sometimes it's the most truthful answer, right?). And, trying to model or explain after the app has been deployed is not only backwards, but hard to accomplish (and seldom done at that). A picture is worth a thousand words... or lines of code.

Another critical visual resource we use in our development process is the wireframe. This extremely simple "drawing" indicates functionality and shows how a user will interact with a system. We use wireframes in addition to use cases because they quickly demonstrate the system flow to our clients, and their simplicity forces us to focus on the matter at hand (what a user does at this point). We refrain from adding any design to wireframes, because design is distracting, so we can spend some good time focusing on the design after we settle on how the system actually works.

Several years ago, I worked with a team that used linked-together html (non-functional but completely designed web pages) to "mock" the site functionality in the manner that wireframes do. This was a high-cost endeavor (it's much simpler to make quick adjustments in a program like Balsamiq than to change an entire html-only mock web app) that kept the entire development team busy – not on designing the software or prototyping the hard parts, but on maintaining throw-away html code. This, and they had to contend with constant comments about the colors and the way the dropdowns worked (This is important stuff, and should be focused on separately.).

Wireframes can be straightforward – the below wireframe expresses the early vision of a project we did for our clients at the Newborn Screening Translational Research Network. Compare the wireframe to the final result at www.nbstrn.org.


This is a very basic description of our process (and the use of bullet points is a visual trick too):

• Talk
• Write down what we discussed (use case and/or technical needs/constraints)
• Draw it (model, wireframe)
• Show and talk and repeat to refine
• Implement

The process can take hours or weeks, but the key final use of the visual in our process is to implement – build what we just discussed in small functional chunks so that our clients constantly see what they asked for in action. This allows them to respond quickly, and for us to adjust rapidly. As a development process, it can be challenging for some clients, because it requires that they be involved in the software development process every step of the way. Our clients pay us to deliver something they don’t have, and the way we do that is to SHOW THEM WHAT WE'RE DOING.

Or, to quote a particularly appropriate proverb, "Tell me and I'll forget; show me and I may remember; involve me and I'll understand."

Of course we’re proud of our process, and visual communication is "just" another tool we use. None of these are novel concepts. Nor, really, are the new FDA labels expressing anything new. People have known for decades years that smoking is the US’s leading cause of preventable death. Here's hoping that a constant visual reminder, on the cigarette packs themselves, will help make some people think twice, or change entirely. In this blog I’ve talked about words and I’ve talked about images. Now, here’s a number for you: 1-800-784-8669. Also known as 1-800-QUIT-NOW. Or maybe this...?


And hey, when's the last time there was this much media attention on cigarette smoking? In the very act of a sneak preview, the FDA has gotten powerful results from this campaign already. And if you think modeling is old-fashioned and time-consuming, because it's much easier to just start writing code, or that html mockups work as requirements docs, or that I've missed a thread, or that it's your right to keep smoking... comment away and let's get a conversation started!

Inspired by Will’s recent book review of Biopunk and the analysis crowd sourcing of the European E. Coli outbreak, I thought I would take another look at DIY (Do It Yourself) biology in this week’s post. Unlike some, I have no interest in trying to run molecular biology experiments out of my kitchen. As anyone who has had the misfortune of trying my cooking would tell you, if there is a way to make a PCR reaction dry and tasteless I'm sure I'd find it. DIY bioinformatics I find more intriguing. I'm not a practitioner as I'm too busy with PSTDIFY (Pay Somebody To Do It For You) bioinformatics, but I like the vision of the lone, amateur scientist, sitting amongst a pile of empty pizza boxes and Red Bull cans finding unknown biological treasure with just their laptop, curiosity and some serendipity.

This vision is not so unlikely. Large biological data sets are readily available including thousands of microarray experiments, genotypes and even full genomes. Someone modestly adept at programming or a package like R can interrogate, correlate and mine this data - and indeed this is happening all the time. What about the true amateur, however, who even lacks programming skills? Can the Excel Warrior or my web savy grandma participate in their own DIY bioinformatics adventure? That’s what I set out to discover this week.

As a test, I went back to a favorite paper of mine by Majewski and Ott ( Genome Research, 2004). What I like about the paper is the number of insights made simply through careful mining of genomic databases. For example, even with inherently noisy data sets like dbSNP and the annotated human genome, the authors were able to clearly see the extent and importance (for splice regulation) of sites near exon-intron boundaries simply by looking at the overall frequency of SNPs discovered in these positions compared to other sites. This figure (F2) from the paper shows the low SNP frequency in the immediate 5' and 3' positions of the intron where it meets the neighboring exons.

My test was to see if I could reproduce at least a part of this analysis by simply using free public tools and without programming. I settled on the web-based analysis tool Galaxy as it seemed to have a lot of the functionality I would need and I wasn’t very familiar with it - making me a better stand in for the Red Bull-intoxicated amateur scientist. After some time poking around, I settled on these steps in Galaxy:

1. Get introns from chromosome 12 via UCSC’s Table Browser (I just did chromosome 12 to keep my data sets manageable for this example).
2. Get all SNPs from chromosome 12
3. Join the introns and SNPs producing a table of only those SNPs that fall within an intron
4. Calculate the position of the SNP relative to the 5’ end of the intron
5. Count up number of SNPs found at each 5’ position
6. Sort results by position (probably not necessary)
7. Limit results to just positions within 50 bp of the exon-intron boundary
8. Plot the SNP frequency vs SNP location

For more detail, you can see my workflow here. And this was my final result:

This corresponds pretty well to the left portion of Majewski and Ott's own intron plot and was finished before cracking my second can of Red Bull. Score one for DIY!

Before I quit my day (and night) job to make room for the waves of empowered amateurs, it's worth pointing out a few minor details. First, this rudimentary analysis glosses over many important details (such as normalizing for intron lengths) and any publication ready workflow would be much more complex. Second, like all tools of this type, Galaxy walks a fine line to balance functionality and usability. It took me quite a bit of exploring to find the right functions and many of these functions probably only made sense to me because I knew a lot about programming, databases, genomics, etc. Third, it's near impossible to match the power, speed and flexibility a programmer has to analyze data with a web based tool like Galaxy. And finally, although I am empowered by Galaxy to do the steps, the know-how of what questions to ask and the science to understand the observations I make comes from many years of experience - unfortunately there's no short cut around that.

With that said, Galaxy has some very nice features and is a powerful addition to the DIY's tool box. Stock up on your Red Bull now.

I often find myself feeling conflicted over medical topics, including personal wellness. Working in this field, beside some of the brightest minds who are using some of the most advanced techniques to drive progress in this area, I can't help but see the value in research and development in the life sciences. There is obviously some questionable, even "bad" science out there and that's not where I'm going with this. Even if the tried and true solve is good, I still can't help but consider an alternative means to fixing whatever is broken. With the speed of research and development of medicine these days, there seems to be a new (and not so new) chemical answer for much of it. I can't always bring myself to just accept the conventional way of remedying my ailment. For example, there have been many times when I choose to try and wait out a headache instead of reaching for the Advil, or when I have employed holistic remedies. When it comes to the controversial topic of vaccinations, I acknowledge that the practice has eliminated (or practically eliminated) many serious diseases. Yet to some degree I can't completely ignore the negative information out there about vaccines, nor the sheer amount and variety of them that now exist and are being mandated or strongly encouraged by medical professionals, school systems, and even society in general.

Based on my philosophical struggles (and those of my peers and friends) it seems with all the scientific progress that has happened, people are now practically regressing to some degree, and are becoming skeptical of much of that advancement itself. According to an article in ScienceDaily (based on a journal paper from the Archives of Internal Medicine), about a third of Americans employ some hybrid form of complementary and alternative medicine. People are seriously concerned about what is going on in the environment and inside of their bodies, and have been smothered with such a wealth of information via the Internet that it makes what was once a simple decision making process (to defer to the recommendation of the physician--period) multi-faceted and quite complicated. The growth in popularity of homeopathic treatments, organic diets, and yoga are just some of the obvious, maybe even trendy, displays of going back to basics.

On a daily basis, I personally try to make responsible decisions, attempting to get and be healthier, but there is no way that I could or would claim that everything can be fixed with some vitamins, breathing exercises, and organic heirloom tomatoes (although they are so delicious, I'm tempted to call them magical). For me--and apparently a third of Americans--there is definitely a delicate balance that needs to be reached in medicine and health. I think personalized medicine will help deliver that balance for many. It's no secret that here at 5AM we are big fans of personalized medicine, and between the costs and time requirements of sequencing dropping, developments in diagnostics (Cancer Dx, Microbiome Sequencing) and the widening understanding that the universal drug is no longer the best way to go, it feels so close! Through personalized approaches, I think individually we will begin to feel less polluted by whatever extras are in the "super-drug," that less things will get broken while some are getting fixed, and that faster results with less side effects will soon be produced. With the onset of personalized medicine, it will hopefully stop feeling like there is conflict between deciding to take what might cure you, and fearing that it will make something else worse. I am professionally and personally fulfilled to be a part of a results-oriented team that not only believes in doing good work, whether it's creating powerful and reliable tools or making sense of data, but also helping others become able to deliver exponentially faster advances too. Efficient systems and solutions that produce quality research results will only push us forward even faster toward more accurate wellness solutions for individual patients, improving the human condition.

I was excited to read this book since I am very interested in bioinformatics and punk (although the book has nothing to do with punk music). Although I wouldn't call myself a practitioner of do-it-yourself (DIY) biology, I do work for a very entrepreneurial bioinformatics and software company. The general theme of Biopunk, by Marcus Wohlsen, is that we are arguably reaching a point in biotechnology similar to where computing technology was in the 1970's. That is, where the germ of successful companies can grow out of innovations by a handful of people working on a shoestring in garages and basements. Think about Steve Jobs and Steve Wozniak starting Apple in their garage, or Bill Gates and Paul Allen starting Microsoft while barely 20 years old. The point, echoed by many of the people who show up in the book, is that until recently biotechnological innovation has been only accessible to scientists at commercial companies or in academic labs.


Wohlsen introduces the reader to some really interesting characters who want to get biological innovation out of the traditional places. There is a 23-year-old who, upon learning that her father has an inherited blood disorder, creates a genetic test for it in her kitchen so she can see if she has the associated genotype, too (she does). There are a couple of guys creating a do-it-yourself PCR testing machine they can take into the South American jungle to test for infectious agents that are affecting the local population. There is a bioinformatics professional who tests his own blood (albeit through an established lab) to see how well folic acid might help him ward off a disorder that causes blindness (for which his 23andMe test showed he had a high risk).

I'll admit to being unsure about how far we are away from people in garages doing genetic engineering and competing with established companies and academics. Wohlsen presents both sides well, although he pretty clearly favors the do-it-yourself crew. My personal guess is that we are still far from understanding how our genomes are interpreted to create whole organisms. It's like having the source code for a piece of software but not a compiler or debugger. I'll bet we are decades away from people, whether they are in sophisticated labs or garages, creating genetically new organisms by anything other than essentially trial and error. Regardless, the stories in the book are thought-provoking and really show what ingenuity people are bringing to bear on biological problems.

However, there is are a few too many hyperbolic statements and other quirks in this book for me to strongly endorse it. There are trivial issues, like when he drops a reference to the classic sci-fi move Blade Runner, saying "synthetic humans shop seedy storefonts for new body parts." I'm pretty sure that never happens in the movie (the replicants go to labs making body parts, but only to get information). There are amped-up and somewhat misleading statements about evolution: "After all, even God has one [a PCR machine] in his toolbox" and "... somehow leapfrog the sheer ingenuity of natural selection ..." The whole point of natural selection if that it doesn't have to be ingenious; the process selects for the best random changes. There are scientific inaccuracies, such as when he refers to genomewide association studies as looking in "areas where researchers had some reason to believe the trait and the gene would be linked" when the whole point of such studies is that they don't pre-identify any specific regions of the genome. And he writes that "strawberries are stuffed with DNA" because they have eight copies of their 200 megabase chromosomes while humans have two copies of three gigabase chromosomes (six billion base pairs in a human cell is more than 1.6 billion base pairs in a strawberry cell).

There are also some places where the organization of the book needs a little work. He tells of a New Zealand academic researcher who dies of being infected with a bacteria she was researching, but ends the chapter with it and doesn't draw any conclusions. It's an academic lab, not a DIY one, so not clear how the story is relevant. He also tells a story of a woman demonstrating extracting plasmids from yogurt-making bacteria at a DIY bio event, but never tells us how the experiment ends up and how she demonstrates the results. The chapter leaves her before she finishes and ends with some DIY bio generalisms ("... if innovation emerges from the ground up, then perhaps its benefits will come more quickly to those at the bottom"). I had written this book, I would have asked the woman who tested herself for her own genetic traits if she had tested it on other people as negative and positive controls (like her father) and if she had repeated the test to make sure she got the same results every time.

Maybe I'm being too critical of stylistic and technical issues, but there are other writers, such as James Gleick, Richard Dawkins, and Matt Ridley, who can take complex scientific topics and turn them into books that hang together thematically and tell complete stories.

Biopunk has tons of good raw material and no end of interesting stories (bridges made of trees, a lab engineering bacteria to produce malaria medicine, etc.) but not all of them seem relevant to the main thrust of the book (that malaria lab is a commercial ones, not DIY). If Wohlsen had expanded his book to cover more territory and taken more care to weave these stories together, it could have been a powerful summary of current trends and future predictions in biotechnology.