Sunday, October 5, 2008

The DNA Network

The DNA Network

Translating “Gene Sherpa” to “Silicon Valley” [Think Gene]

Posted: 05 Oct 2008 08:45 PM CDT

Dr. Steven Murphy, “The Gene Sherpa” has nagging me to write this…

Steve is upset that DTC genomic startups, specifically, 23andMe, aren’t following “protocol.” He listed a bunch of institutional policies and standard procedures that I don’t care to remember in mind-dump blog post he’s since removed, and used these “violations” as justification to decry 23andMe and friends as reckless and unethical.

So I’m in New York City with Steve, and I say: “listen, Steve, the first thing anyone in Silicon Valley will think when you prescribe a stack verbose regulations will be ‘how can I ignore this?’ These people, who are the same people of scientific marvels like Google, have a dim opinion of all bureaucratic authority —including medical authority— and for good reason. Nobody there cares what some document says, they only care why those legacy policies exist and what problems they’re supposed to solve. Why? Because they think that they can solve those problems better. And you know what? They probably can.

“So, if you want to help —if you truly want to do good— rather than rant and be willfully ignored by your targets, then you have to simply state the problems, why they’re problems, and suggest a simple, non-authoritative, actionable way to solve them.”

(Also, Steve doesn’t want to provide free consulting to Coriell’s free genomic test’s competitors because he’s probably going to have a special surprise for the people of New York, but, you know, deals can fall through, so I’ll wait until the contract is signed and all that to tell you anything.)

So, here they are:

1) Transparency by Independent Expert Oversight

Problem: If a company acts unethically, how will public know about it, and how will the public know when it’s fixed?

Somebody must be both empowered and responsible for the ethical operation of a company obviously and independently.  That way, when we hear no reports of problems, it’s because we know the company is operating ethically, not because there’s no way to report problems.

Question: Who in DTC genomics is responsible and empowered to report problems?

Why This Problem Matters: Genomic testing companies are a new blend of software, health care, and private medical records. Very few ethical issues matter in information technology itself, but health and medicine are especially fraught with ethical concerns. It’s necessarily yet not sufficient to “don’t be evil.” The greater the ethical risk, the greater confidence the public must have in the company, and the greater the transparency must be.

Solution: Appoint a small board of diverse experts appointed to passively oversee the the company’s operations and report problems to the public in a standard and obvious way.

Is this a necessary expense? No: it’s unlikely to appeal to the press and general public and generate immediate sales. But it does improve the long-term company image as seen by the informed and the medical research establishment. I notice that 23andMe is hiring a durrbizdev. New bizdev director, this board and the act of forming it would be the single most effective “business development” initiative for these people. Throw in some soulful black preacherman, some Berkley femnazi community leader, and a some whiny tech author in glasses and you’ll have yourself a SWPL media event too, ripe for plucking off the press release vine.

A similar board may already exist, but that it’s private, poorly publicized, or not well defined. Define it, publicize it, spice it up as noted above, problem solved —plus one more guaranteed NYTs article.

Existing Policy: The existing policy is to form an Institutional Review Board (IRB). Steve could certainly speak better about this, but the end result should be implementation of my simple statement solution.

Is it a necessary expense to use existing policies to form an official IRB? My suspicion is that existing policies are painfully slow, abstruse, and expensive, but existing institutions will to trust you more if you use them. They’d be suspicious if you created your own policy, even if your policy is better, because usually when people make their own policy, it’s in their own best interest, and because people trust what they know irrationally. So, can you convince people that matter that your solution is as good or better as an official IRB and deserves as much or more trust? Is the difference between your solution and an IRB less than the extra convincing?

My general suggestion is to first solve the problem rationally, and then have smart people justify your solution with the IRB official procedures. That’s because trying to follow any government document to solve anything without already knowing what you’re trying to do is a good way to make something asinine.

2) Opting Out

Problem: How can people leave the system?

If people cannot terminate their inclusion in an ongoing medical study, then what incentives keep that study to perpetually act in the best interest of its members, and what recourse do members have if disapprove of the company?

Note this except from the 23andMe Terms of Service:

Your saliva, once submitted to and analyzed by us, becomes our property. Any genetic information derived from your saliva remains your information. We retain the rights set forth in the consent form and any additional terms of service.

and from the Consent Agreement:

You have the right to delete your genetic information from our systems. Within thirty (30) days of receiving your written request, we will delete your account, and your information will not be included in any future research, including future research by other organizations. Any research conducted prior to the end of the thirty (30) day period following receipt of your request will not be altered or halted. Once your account is deleted it will not be retrievable.

Once information is shared with research partners, we cannot guarantee that it will be destroyed upon request.

Who cares about the saliva sample? It’s a legal convenience and standard procedure that physical objects you mail to a company become property of that company. What’s important is the genomic information is deleted to the best of their power on request, and that is contractually promised.

But clearly, there is confusion that’s making a problem, because:

  1. If Steve was upset, and he’s a medical doctor of genetics and a smart, informed, active participant in the medical genomics community, then it’s not obvious enough that people can leave the system.
  2. Keeping the saliva sample violates some medical research convention, and that offends the ethical standards of others in the medical research community.

Question: If a participant no longer wants to participate in a DTC genomic service, for personal reasons or as protest, what recourse do they have?

Existing Policy: Destroy the biological sample (the saliva) and the data on request. There’s probably some byzantine government policy to do this that Steve would know about, but I don’t care to look up.

Solution: Add account deletion to the main FAQ and promise to destroy the saliva sample.

First, know who this solution is meant to appease: the existing medical research and genomic medicine establishment. If it helps, think of them as old grouchy prigs sitting in ivory towers, and when something procedural is awry— regardless of its superficiality —their buttholes get very tight and they are unable to say nice things about your service. You don’t want to do that to the elderly, do you?

The addition to the FAQ is easy. Even better, add links to anchors to the relevant excepts in the consent form and terms of service in the FAQ text. As a bonus, this should help comfort customers, too.

I’m not suggesting anything else besides a promise to destroy the sample. It can still be your property, and you don’t have to implement any official policy. You can do that later. But in the meantime, this is an easy win that helps your image in the medical research community.


Predictably, 23andMe is crushing its competitors in everything information: software, marketing, web product, because that’s its founding background. Likewise, Coriell is crushing its competitors in everything medical research: policy, ethics, funding, because that’s its founding background. Specifically, everybody knows about 23andMe and their website and PR is excellent, but Coriell provides free, medical testing. Yes, it’s not “for education and research” like 23andMe. It’s a certified medical test. But, hardly anyone knows about Coriell, and their web service and marketing ranges from “non-existant” to “sucks.”

Imagine if West Coast 23andMe and East Coast Coriell joined forces to create a service with both strengths. Now that would be a hybrid genetics experiment of note!

Media-Induced Confusion about LRRK2 Discussed in 23andMe Forums [Think Gene]

Posted: 05 Oct 2008 04:18 PM CDT

Confusion about g2019s, LRRK2, and Parkinson’s Disease still circulates after the vacuous media blitz about Sergey’s blog. Here’s an except of a conversation from the 23andMe private member forums.

Not included is Paul Wick’s note: “I guess the only question in my mind for 23andMe from a regulatory perspective is at what point this is diagnostic vs educational; I know it's a different kettle of legal fish in the latter case. ”

Indeed. However, g2019s is not used to make nor confirm clinical diagnosis of Parkinson’s Disease, so I think this still counts as “education.” (I so hate that word.) Personally? Assuming 23andMe makes the appropriate medical care available, I’d love to see them report diagnostic mutations.

PaulWicks asks:
LRRK2 mutations?

A patient with Parkinson's Disease (PD) posted today on the PD board of PatientsLikeMe that Google founder Sergey Brin has tested positive for LRRK2, a mutation that is implicated in some cases of inherited PD. See the original blog here: http://too.blogspot.­com/­When I look at my raw data I can see some listings next to LRRK2 but I need some more help to interpret that data; how can I tell what is / isn't pathogenic?

Paul Wicks
R&D Director @

Andrew Yates
Andrew Yates responds:
LRRK2 mutations?

Hey Paul,

First, see http://www.helixgene­.org/press/g2019s­ for a simple fast-fact press release about the g2019s mutation and Parkinson’s Disease.

“tested positive for LRRK2″ is a misnomer, what you mean is “tested positive for a variation of LRRK2.” The distinction is important, bare with me:

LRRK2 is a gene that everybody has. Everybody would test positive for LRRK2.

Actually, the test is for a single SNP[1] in the gene LRRK2 called g2019s. I see you’re R&D;, so in computer code, the test evaluates:

if (YourGenome[39020469][0] == A        || YourGenome[39020469][1] == A)

What does this mean? LRRK2 describes the protein “dardarin.” Think of a protein as a string of amino acids. This “bit flip” in your DNA changes the amino acid glycine at position 2019 to amino acid serine in dardarin. As code, this is like:

if (YourGenome[39020469][0] == A        || YourGenome[39020469][1] == A) then     dardarin[2019] = serine else     dardarin[2019] = glycine endif

Note: g2019s is an abbreviation for “replace (g)lycine @ position 2019 with (s)erine”

What does THIS mean? It’s thought that g2019s mutated dardarin increases the protein MAPK’s activity, and this gain of function mutation leads to neuron death. That’s why g2019s is autosomal dominant: you only need one copy of the mutation to increase function and to create a toxic effect.[2]

This is only a test for one very specific change in LRRK2. There are at least 30 known SNPs leading to 20 known amino acid substitutions mutations for LRRK2, though most of these have not been as well studied or are as medically relevant as g2019s. But a mutation can be be anywhere on the gene and cause all sorts of subtle and complex effects depending on other proteins and the environment…


So, as you can see, this is a very complex system with many minutia that aren’t well understood. While I thought Sergey’s post was quite good, it’s _extremely irritating_ to see this sloppily reported in the media (like the NYTs) leading to significant medical misinformation at worst and general confusion at best. This is complex and important medical information, and if a NYT tech reporter[3] with a comp sci degree from Stanford completely bungles elementary bayesian reporting because he copy-pasted the first percentage out of a press release and link cited a blog so he could go home early for the weekend… then we have a lot of work to do. Why not just read “MSN Health’s “Top Ten Mutations That Could Change Your Life?” Enraging.

I can tell you that the genomic medical community is PISSED, and it’s not even 23andMe’s fault. As far as I can tell, 23andMe’s g2019s test and report are very good. But, the medical community only knows what they’ve read in the NYTs because they don’t have 23andMe accounts. All they know is “Holy Crap! the NYTs says that 23andMe says that you could have an up to an 80% risk of Parkinson’s disease! Those huckster criminals!” Clearly, that’s not what 23andMe or Sergey’s “Too” actually says, but that’s the cost of standard of media reporting about genomics today.

[1] SNP, or single nucleotide polymorphism. “single nucleotide” means one DNA letter is changed, and “polymorphism” means a known variant that’s found in at least 1% of a population.

[2] It’s thought that this mutation makes dardarin more active because serine can be phosphorylated and glycine can’t be, and this helps the dardarin change shape such that ATP can better access its active site. Dardarin adds phosphate to other proteins including the protein MAPK, and this in turn makes MAPK too active. This is not my expertise, so any corrections or elaborations are welcome.

[3] Miguel Helft and his editor, Damon Darlin

Adaptationist Junk [evolgen]

Posted: 05 Oct 2008 01:00 PM CDT

The human genome (like all mammalian genomes) is loaded with sequences that don't perform any known function. And many of these sequences are junk. And it's not just mammals -- many animal genomes are loaded with junk, as are those of other eukaryotes. That's not to say that some of the sequences of unknown function do, in fact, have a function that we have yet to identify. However, much of the junk comes from the remnants of transposable element (parasitic sequences that hop around the genome).

That's not the impression you get from the introduction to a review of transposable elements (TEs) published in the most recent issue of Cell (doi:10.1016/j.cell.2008.09.022):

Following the discovery of transposons in plants and bacteria, the presence of mobile DNA in eukaryotic species gained widespread acceptance. However, the concept of "controlling elements" gave way to disparaging terms such as selfish DNA and "junk DNA." Nevertheless, the notion of transposable elements as merely molecular parasites, benign at best and powerful mutagens at worst, that hijack cellular mechanisms for their own selfish propagation, seemed incomplete to some biologists. Given that evolution tends to dispose of that which is useless and harmful for a species, it was curious that the genome should be cluttered with so much "junk." Now we understand that genomes have coevolved with their transposable elements, devising strategies to prevent them from running amok while coopting function from their presence. Repetitive DNA, and retrotransposons in particular, can drive genome evolution and alter gene expression. Evolution has been adept at turning some "junk" into treasure.

I don't see how the hypothesis that transposons are selfish genetic elements conflicts with the observation that host genomes have coevolved with these parasites. That is, in fact, what we expect to happen between hosts and parasites -- red queen and all.

Read the rest of this post... | Read the comments on this post...

How much is that in Apollos? [Mailund on the Internet]

Posted: 05 Oct 2008 08:06 AM CDT

I just have to share this comparison of the finacial crisis bailout and large-scale scientific programs:

… For example, you could build 70 Large Hadron Colliders (or maybe just one Ultra-Humongous Hadron Collider).

By the way, the entire US space effort through Apollo was about 100 billion in current dollars,  so we could do seven Apollos for the amount of the bailout.

You know, a few hundred billions here and a fwe hundred billions there, all of a sudden we’re talking real money!

Webicina will be open to patients as well: 2 days left [ScienceRoll]

Posted: 05 Oct 2008 04:45 AM CDT

We are working hard to get Webicina ready by the 7th of October. This Tuesday, after months of preparation, we will launch it officially. We decided to open the service for patients as well. That is why it’s easy to say now Webicina is really trying to build a bridge between e-patients and physicians.

The services we will provide:

  • Consulting: presentations about web 2.0 and medicine in person; or online (webinars, Second Life workshops).
  • Online Image Building Solutions: e-mail support for 3 months about the tools and methods medical professionals need to build a proper online reputation.
  • E-Courses: materials and step-by-step tutorials through which users can learn to use the web 2.0 tools and methods they need. Premium account provides a 3 months long access and e-mail support, while the Basic account does not include e-mail support.
  • Personalized Medicine 2.0 Package: a personalized set of web 2.0 tools designed to solve the problems of physicians, health care workers and patients. Membership includes one extended e-mail containing all the web 2.0 tools, services and information dedicated to a specific medical condition or a medical specialty.

Please join the discussion about how to change the way medicine is practiced and healthcare is delivered from the 7th of October!


ScienceStage: Knowledge Transfer [ScienceRoll]

Posted: 05 Oct 2008 03:29 AM CDT

There is a new scientific community site, the newest addition to my list, which aims to help scientists from the different fields of science to connect and share.

The universal online portal for science, advanced teaching and academic research. is a virtual conference room, lecture hall, laboratory, library, and meeting venue all in one. It offers new methods of scientific presentation, scientific discourse and academic knowledge transfer by using video streaming, audio streaming and text features.

There are more and more similar sites, but just a few of them will survive…


Blogging about blogging [business|bytes|genes|molecules]

Posted: 05 Oct 2008 01:22 AM CDT

I don’t often write about blogging. Seems somewhat circular to do so, but there has been a lot of discussion on the web lately about what blogging really means and what the role is, etc. Having been at it for a bit, and hopefully with some visibility, perhaps I can pen some thoughts down on this subject.

In an age when the definition of a blog and who is a blogger gets more and more cloudy, I still consider myself a blogger, although that word brings back memories of an elegant medium for a more civilized age. To be a blog is just one format for making the voice that all of us have to a degree more public. There are many others, but this is the one that makes the most sense when you want to write about current events, the things you have opinions about and when you want to rave and rant. The chronological format, the ability to comment and have a “conversation” with others makes blogs very convenient in a number of ways.

TIMTOWDI. Every Perl programmer knows what that means. It also applies to pretty much every third thing out there, including blogging. There is no one way to describe blogging, what constitutes a blog, what constitutes a science blog, what the role of a blog should be, etc. For me a blog is very personal, for others it’s a way of communicating with users of a particular piece of code, or publishing news about your products, and increasingly the modern op-ed. You can write about visualization of publications, add additional analysis to a paper, introduce people in your team, or talk about your mission statement. Some bloggers are journalists, and some journalists are bloggers, but we are always trying to talk in generalities when it comes to that subject.

So here is a simple request. Can we stop trying to paint everything in the broadest brushstrokes possible? All you need to be a blogger is the ability (or lack thereof) to write, and an appropriate CMS that allows you to express yourself in the appropriate format. And that’s just one opinion. YMMV

Image from xkcd

Reblog this post [with Zemanta]

No comments: