Four short links: 5 August 2009

Rebooting Britain, Revealing Errors, Reproducing Generators, Netflix Culture

1. Reboot Britain Video Archive -- video from the talks at Reboot Britain are online. The event also produced a essay set (PDF), CC-licensed. (via Paul Reynolds)
2. Revealing Errors -- Benjamin Mako Hill blog using computer errors as starting points for understanding how computers control the world around us. (via Dan Meyer)
3. New Microbe Strain Makes More Electricity, Faster -- University of Amherst researchers made current-generating bacteria work harder to live, and in five months had a strain that made an 8x larger current.
4. Netflix Culture -- readable slide deck which talks about the Netflix company culture. It's hard to read it and not nod in full agreement. (via joshua on Delicious)

Four short links: 28 July 2009

UI Library, 3rd Party Wave Server, Mobile Phones + Parasites, Single API to Cloud Providers

1. CNMAT Resource Library -- The CNMAT Resource Library is our fast growing collection of materials, sensors, gestural controllers, interface devices, tools, demos, prototypes and products - all organized and annotated to support the design of physical interaction systems, "new lutherie" and art installations. (via egoodman on Delicious)
2. PyGoWave Server -- first third-party Google Wave server, based on Django.
3. Mobile Phones Identify Parasites and Bacteria -- UCB Researchers developed a cell phone microscope, or CellScope, that not only takes color images of malaria parasites, but of tuberculosis bacteria labeled with fluorescent markers.. The sensor network is built out, and the computers in our pockets surprise us with their uses. (via BoingBoing)
4. libcloud -- a unified interface to cloud providers, written in Python and open source. Covers EC2, EC2-EU, Slicehost, Rackspace, Linode, VPS.net, GoGrid, flexiscale, Eucalyptus. (via joshua on Delicious)

Four short links: 27 July 2009

1. Ignite OSCON -- 56m of video from Ignite OSCON. They're all great, but Dan Meyer remains the highlight for me.
2. gheat -- a maptile server in Python, delivering heatmaps to be superimposed on Google Maps. Handy for visualization fiends.
3. CaDNAno -- open source software for design of 3-dimensional DNA origami. One of George Church's projects. I love the combination of math, biology, and whimsy in open-source giftwrap. (via timoreilly on Twitter)
4. CommentPress -- an open source theme for the WordPress blogging engine that allows readers to comment paragraph by paragraph in the margins of a text. Annotate, gloss, workshop, debate: with CommentPress you can do all of these things on a finer-grained level, turning a document into a conversation. It can be applied to a fixed document (paper/essay/book etc.) or to a running blog. I'm taking a greater interest in tools that channel and focus participation rather than simply providing "edit this page". (via gov2.net.au's issues paper)

Four short links: 15 July 2009

A collection inspired by Science Foo Camp attendees

1. Endogenous steroids and financial risk taking on a London trading floor (PNAS) -- We found that a trader's morning testosterone level predicts his day's profitability. We also found that a trader's cortisol rises with both the variance of his trading results and the volatility of the market. Our results suggest that higher testosterone may contribute to economic return, whereas cortisol is increased by risk. Our results point to a further possibility: testosterone and cortisol are known to have cognitive and behavioral effects, so if the acutely elevated steroids we observed were to persist or increase as volatility rises, they may shift risk preferences and even affect a trader's ability to engage in rational choice.
2. The Origin of Universal Scaling Laws in Biology -- eye-opening paper that blew my mind. Highlight of Sci Foo was meeting the author and shaking his hand. Relates metabolic rate, size, heart rate, and lifespan by applying physics to biology.
3. Ushahidi -- open source software for managing disasters. The Ushahidi Engine is a platform that allows anyone to gather distributed data via SMS, email or web and visualize it on a map or timeline. Our goal is to create the simplest way of aggregating information from the public for use in crisis response.
4. Dissecting the Canon: Visual Subject Co-Popularity Networks in Art Research -- In this paper we analyze a classic da- taset of art research, which collects ancient art and architecture and their Western Renaissance documentation since 1947. [T]here is clearly a long tail of monument popularity.

Four short links: 9 June 2009

Biological Radio, Laggy Smart Grids, API Moneys, and Pubsub Server

1. Drawing Inspiration From Nature To Build A Better Radio -- based on the design of the cochlear, this MIT-built RF chip is faster than others out there, and consumes 1/100th the power. Biomimicry and UWB radio are on our radar.
2. Why the Smart Grid Won’t Have the Innovations of the Internet Any Time Soon -- While it’s significant that utilities are starting to build out smart grid infrastructure, utilities are largely opting for networks that provide connections that are far from real time, and this could stifle the desired innovation. [...] smart meter data that is pushed to Google’s PowerMeter energy tool has to make its way back to the utility before it can be sent to Google. That means that even for Google’s energy tool, there can be both a significant delay before information reaches consumers, and significant gaps in energy data details. These delays and gaps can undercut the premise of how smart meter technologies will empower consumers to make decisions about their energy use based on real-time costs. Smart grids (houses and devices able to take use of instantaneous pricing changes) have the potential to help us with our energy obesity problem, but the architecture must be right.
3. API Value Creation, Not Monetization -- On the side of the unexpected but interesting outcomes, Kevin said they have seen a flurry of internally developed business applications. In the past many valuable, internal-facing projects were turned down because the programs had to meet strict top line to bottom line ratios. With the availability to data and services, many teams within the company now have access to things they didn’t in the past, and project costs have been minimized. Throughout the company, consumers of the API have been able to launch successful projects that have created additional revenue and have reduced the overall development costs for new projects. Some solid numbers and names to help convince businesses to offer APIs, though the battle is still much harder than it should be.
4. Watercoolr -- a pubsub server for your apps. A channel is a list of URLs to be notified whenever a message is posted to that channel. Clever little piece of infrastructure for web apps, embodying the Unix philosophy of small tools that each do one thing very well. (via straup on Delicious)

Four short links: 12 May 2009

Storage Superfluity, Data-Driven Design, Twit-Mapping, and DIY Biohacking

1. Lacie 10TB Storage -- for what used to be the price of a good computer, you can now buy 10TB of storage. Storage on sale goes for less than $100 a terabyte. This obviously promotes collecting, hoarding, packratting, and the search technology necessary to find what you've stashed away. Analogies to be drawn between McMansions full of Chinese-made crap and terabyte drive full of downloaded crap. Do we need to keep it? Are there psychological consequences to clutter? (via gizmodo)
2. In Defense of Data-Driven Design -- a thoughtful response to the "Google hates design!" hashmob formed around designer Douglas Bowman's departure from Google. When you’ve got the enormous traffic necessary to work out if miniscule changes have some minor, statistically significant effect, then sure, if you can do it quickly, why wouldn’t you? But that’s optimization that should happen at the very end of the design cycle. The cart goes after the horse. Put it the other way ‘round and you have a broken setup. It doesn’t mean horses suck. It doesn’t mean carts suck. Carts are not the enemy of horses. Optimization is not the enemy of design. Get them in the right order and you have something really useful. Get them the wrong way around and you have something broken.
3. Just Landed: Processing + Twitter + Metacarta + Hidden Data -- Jer searched Twitter for "just landed in", used Metacarta to extract the locations mentioned, and then used Processing to build visualizations.
4. Do It Yourself Genetic Sleuthing -- MIT is starting a hotbed of DIY biologists. The 23-year-old MIT graduate uses tools that fit neatly next to her shoe rack. There is a vintage thermal cycler she uses to alternately heat and cool snippets of DNA, a high-voltage power supply scored on eBay, and chemicals stored in the freezer in a box that had once held vegan "bacon" strips. Aull is on a quirky journey of self-discovery for the genetics age, seeking the footprint of a disease that can be fatal but is easily treated if identified. But her quest also raises a broader question: If hobbyists working on computers in their garages can create companies such as Apple, could genetics follow suit? It's unclear what those DIY-started "genetics" companies would look like--the potential is there, but it's yet to met the right problem. (via Andy Oram)

Just Landed - 36 Hours from blprnt on Vimeo.

Four short links: 5 May 2009

Spies, Community, International Success, and DNA Origami

1. Supermap -- The CIA's venture capital arm, In-Q-Tel, is paying an undisclosed sum to California-based Geosemble Technologies to develop an intelligence version of the "geospatial data integration and layering technology" that the company developed for use by urban planners, real estate investors and market analysts. The technology combines overhead imagery, maps and heavy-duty data mining to create a map-based intelligence capability reminiscent of the Pentagon's former Total Information Awareness program. When the project is done - and In-Q-Tel won't say how soon that might be - CIA agents will be able to merge aerial images and electronic maps on a computer screen. Then they will be able to click on the building or other item of interest and all manner of information will pop up: who the tenants are, phone numbers, company records, links to company and organization Web sites, news reports related to the tenants or incidents at the address, property records, tax data and more. I love that Cheap Suit Susan, your local real estate agent, had the technology before the CIA. It's like learning that Lionel Hutz has a missile defense system to stop his house being TPed.
2. 7 Harsh Truths About Running Communities -- As the leader of your community, your personality sets the tone. As a result if the community behaves in ways you do not want, then you only have yourself to blame. I have seen many bloggers write about the negative comments they get on their posts. In most cases this is due to the tone they themselves strike in their writing. Although there are exceptions I believe that users will respond in the same voice you yourself set. If you are irreverent, then so will your users be. If you are rude, expect rude responses. "Social software" is an anachronism-software that doesn't let users interact has become antisocial software. Every web creator needs to know what successful communities have in common. (via Julie Starr)
3. Lingopal is Big in Japan (Lance Wiggs) -- Turns out we are biggest in Japan. We have done no marketing there - it is all organic growth as our google ad writing and PR ability is not so good in Japanese. More anecdata for my belief that, while chance favours the prepared mind (as Louis Pasteur said), we routinely use post-hoc rationalisation to explain why it was inevitable that this or that lucky SOB hit it big.
4. DNA Origami Seeds: Bottom-Up Methods for Molecular Self-Assembly (US News) -- Winfree's coworker at Caltech, Paul W. K. Rothemund, pioneered the seed-DNA technology that allows tiny "DNA origami" structures to self-assemble into nearly arbitrary shapes (such as a smiley face and a map of the Western Hemisphere). The researchers designed several different versions of a DNA origami rectangle, 95 by 75 nanometers, which served as the seeds for the growth of different types of ribbon-like DNA crystals. The seeds were combined in a test tube with other bits of DNA, called "tiles," heated, and then cooled slowly. At the lower temperature, the tiles start to stick to each other and to the origami. In this way, the DNA ribbons self-assemble, but only into forms such as ribbons with particular widths and ribbons with stripe patterns prescribed by the original seed.

Four short links: 4 May 2009

Maps, Africa, Protein, and Rockets

1. Old Japanese Maps on Google Earth Unveil Secrets -- Google criticised for putting up map layers showing the towns where a discriminated-against class came from, because that class is still discriminated against and Google didn't put any "cultural context" around it. Google and their maps didn't make the underclass, Japanese society did. Because they're sensitive about having the problem, they redirect their embarrassment into anger at Google. You could make a long and profitable career in IT consulting simply by charging to say "it's not a technical problem" and you'd be right more times than wrong.
2. See Africa Differently -- using the Internet to reframe a continent. Videos, essays, and more, all designed to get you seeing the majority of Africa, which isn't defined by conflict and famine. (via NY Times book review)
3. Fold.it - Solve Puzzles for Science -- science harnesses our "cognitive surplus" by inviting us to help solve the problem of protein folding, one of the hardest in biology. (via auckland_museum on Twitter)
4. Arduino Telemetry Payload in Class C Rocket (Jon Oxer) -- Because class-C rockets are so small and light they can't lift much of a payload and I had to keep the mass of the electronics as small as possible. You can get a sense of scale from this photo which shows a small white bundle in the bottom of the nosecone. Inside that bundle is an Arduino Pro Mini 5V/16Mhz, a 433Mhz transmitter module, and a Lilypad 3-axis accelerometer. PCBs ... in ... Spaaaaace!

Four short links: 23 Apr 2009

Multitouch, visualizations, body hacks, and ubicomp:

1. Dell Demos Multitouch on the Studio One 19 (Engadget) -- the multitouch software on this baby is Fingertapps from the New Zealand company Unlimited Realities, whose founder was at Kiwi Foo Camp this year. Multitouch hits consumer PCs in a very mainstream way.
2. Circos -- open source Perl library to produce beautiful circular data displays. (via flowing data)
   
3. Brain Gain: The Underground World of “Neuroenhancing” Drugs (New Yorker) -- more on the body hacks theme of radical and literal self-improvement, as originally documented by Quinn Norton. What I found interesting was that when BoingBoing linked to it, they quoted the "Provigil might make us smarter" bit, and when MInd Hacks linked to it, they quoted the negative effects of amphetamine-based drugs.
4. Towards the Web of Things: Web Mashups for Embedded Devices -- slides and notes for a presentation given at MEM 2009. Basically saying that the Internet of Things should be built on JSON and REST, with demo. (via Freaklabs)

Four short links: 8 Apr 2009

Bias, RFCs, virus batteries, and a glimpse at life beyond record labels (the last item features profanity, beware):

1. Bias We Can Believe In (Mind Hacks) -- Vaughn asks the tricky question about the current enthusiasm for Behavioural Economics in government: where are the sceptical voices? As he points out, It's perhaps no accident that almost all the articles cite a 2001 study that found that simply making the US's 401(k) retirement savings scheme opt-out instead of opt-in vastly increased participation simply because it's a hassle to change and employees perceive the 'default' as investment advice. But it's probably true to say that this example has been so widely repeated but it's one of the minority of behavioural economics studies that have looked at the relation between the existence of a cognitive bias and real-world economic data from the population. And it's notable that behavioural economists who specialise in making this link, a field they call behavioural macroeconomics, seem absent from the Obama inner circle.
2. How The Internet Got Its Rules (NYTimes) -- about the first RFCs, which became IETF. The early R.F.C.’s ranged from grand visions to mundane details, although the latter quickly became the most common. Less important than the content of those first documents was that they were available free of charge and anyone could write one. Instead of authority-based decision-making, we relied on a process we called “rough consensus and running code.” Everyone was welcome to propose ideas, and if enough people liked it and used it, the design became a standard. (via Glynn Moody)
3. Viruses Could Power Devices (Science News) -- Ions and electrons can move through smaller particles more quickly. But fabricating nano-sized particles of iron phosphate is a difficult and expensive process, the researchers say. So Belcher’s team let the virus do the work. By manipulating a gene of the M13 virus to make the viruses coat themselves in iron phosphate, the researchers created very small iron phosphate particles. (via BoingBoing)
4. Amanda Palmer's Label-Dropping Game -- interesting email from Amanda Palmer to her fans about trying to get dropped from her label. i had to EXPLAIN to the so-called "head of digital media" of roadrunner australia WHAT TWITTER WAS. and his brush-off that "it hasn’t caught on here yet" was ABSURD because the next day i twittered that i was doing an impromptu gathering in a public park and 12 hours later, 150 underage fans - who couldn’t attend the show - showed up to get their records signed. no manager knew! i didn’t even warn or tell her! no agents! no security! no venue! we were in a fucking public park! life is becoming awesome. and then the times they are a-changing fucking dramatically, when pong-twittering with trent reznor means way more to your fan-base/business than whether or not the record is in fucking stores (and in my case, it ain’t in fucking stores).

ETech Preview: Creating Biological Legos

If you've gotten tired of hacking firewalls or cloud computing, maybe it's time to try your hand with DNA. That's what Reshma Shetty is doing with her Doctorate in Biological Engineering from MIT. Apart from her crowning achievement of getting bacteria to smell like mint and bananas, she's also active in the developing field of synthetic biology and has recently helped found a company called Gingko BioWorks which is developing enabling technologies to allow for rapid prototyping of biological systems. She will be giving a talk entitled Real Hackers Program DNA at O'Reilly's Emerging Technology Conference, March 9-12, in San Jose, California. And she's joining us here today. Thank you for taking the time.

RESHMA SHETTY: No problem. Happy to be here.

JAMES TURNER: So first of all, how do you make bacteria smell nice, and why? I get an image of a commercial, "Mary may have necrotizing fasciitis, but at least her hospital room smells minty fresh."

RS: Well, the original inspiration for the project was the fact that for anybody who works in a lab, who works with E. coli, when you grow cultures of the stuff, it just smells really bad. It smells really stinky, basically. And so our thought was, "Hey, why don't we reengineer the smell of E. coli? It'll make the lab smell minty fresh, and it's also a fun project that gets people, who maybe aren't normally excited about biology, interested in it because it's a very tangible thing. I can smell the change I made to this bacteria."

JT: So what was the actual process involved?

RS: So the process was, you basically take a gene, we took a gene from the petunia plant, which normally provides an odor to the flower, and you place that gene into the E. coli cell. And by supplying the cell with an appropriate precursor, you make this minty smell as a result. So it's fairly straightforward.

JT: Your degree, biological engineering, is a new one to me. How is it different from biochemistry or microbiology or genomics or any of the other traditional biotech degrees?

RS: Well, biology and biochemistry, and so on, are concerned with studying the natural world. So I'm going to go out and figure out how the natural world works. Biological engineering, instead, is really all about saying, "Hey, we have this natural world around us. Biology is, in some sense, a new technology through which we can build new engineered biological systems." Right? So the idea is, what's the difference between physics and electrical engineering? Electrical engineers want to go build. So in biological engineering, we're interested in going and building stuff, too. But using biology, rather than physics, as the underlying science of it.

(continue reading)

Four short links: 9 Jan 2009

Four questions, one per link: what next, can it solve a big problem, what's the final boss for Python programming, and why on earth would anyone want yogurt that glows in the dark?

1. End Times - gloomy piece on the future of journalism, to be added to the large pile of other gloomy pieces on the future of journalism (e.g., Bad News, Good News). The critical problem is still how to pay for journalism if the new media revenues are significant lower than old, and if the new media economics decree that journalism is dead then who fills the social good role that journalism's death will leave?
2. Ward Cunningham's Visible Workings - an intriguing glimpse, from March last year, into the way Ward lays out web interactions. Nice system for laying out these interactions, but it's also fascinating for how it makes transparent what will happen as a result of the data you submit. How scalable is this? Could it tackle privacy?
3. Project Euler - fun programming exercises that require more than math to finish. We learn by doing, not by reading, so interesting exercises are part and parcel of training. It's interesting to see educators are moving from being authors to being game designers, providing a series of staged challenges that make us stronger by defeating them. I'm presently dieing in as many ways as I can while learning iterators and generators in Python, as a way of ensuring I have Python's "game physics" sussed.
4. Rise of the Garage Genome Hackers - more on hobbyist molecular biology. It mentions DIYBio, the Cambridge biohacker collective that I first heard about at BioBarCamp. (via Glynn Moody)

Richard Jefferson Interviewed in Com Ciência

I enjoyed this interview with Richard Jefferson (caution: PDF) from Com Ciência No. 102, October 05, 2008. Richard runs CAMBIA, a group that fights for open innovation in biological sciences. He's particularly cautionary about the potential for patents to greatly restrict the development of Synthetic Biology (SB):

But don't doubt there will be some very interesting biological understanding that emerges from engineering experimentation in SB. Probably some very helpful insights. But to make it truly open? It simply won't happen. There will be too many dominating patents that can stop an SB implementation that never even contemplate SB.

Auckland University Bioengineering Institute

I am an industry advisor to the Auckland University Bioengineering Institute and got a tour on Tuesday. It was inspirational! They sprawl over several floors of a tall concrete building in Auckland, expanding from their cramped one-floor presence. Everywhere you look there are people with soldering irons, laptops, and batteries working on devices that sit between hardware and biology.

I've been advising on their Physiome project, which seeks to computationally model human bodies at various levels including genes, proteins, muscles, skeleton, and skin. The idea is to develop a model such that you can "sample" a person's key physical trait, plug it into the model, and predict the details of their body's structure and response. I like to think of it as the OSI 7-layer stack of the human body.

You can see a list of their projects online. I saw the telemetry group's work on powering heart pumps, which assist a failing heart while the patient waits for a transplant. At the moment, heart pump patients have power cables sticking out of their chest and consequently many die from infection before they can receive the transplant. The telemetry group is working on wireless power transfer to the devices. It's the same inductive power ideas that made a splash last month when Intel demoed wireless power devices at their developers forum.

I also saw a lot of very sexy hardware. They have a 3D printer, a laser cutter, and a monstrously heavy metal machining tool that had to be delivered through the window by a crane (and which required multiple engineering checks of the floor's capability to hold it up). All these take designs from CAD diagrams, so the researchers can conceive of a part, design it, and produce it without the lengthy turnaround times and erratic tolerances of traditional machine shops. They're actively expanding their metal shop.

I also saw an artificial muscle in the biomimetics group. It was flapping at an adjustable rate, needing a high-voltage low-current power which can be made quite small and efficient. They were still trying to find an application--candidates they're investigating include moving small devices (the way some organisms use cilia to move around), and using may of them to form a crowd-surfing type of conveyor belt (pictured).

They have programmers, too—modelling is a big part of their work, and every model needs lots of sexy outputs. One of their coders, Duane Malcolm, came to Kiwi Foo earlier this year. He's an open source hacker, with Sparkfun Arduino kits at home, who has been hacking in XUL, RDF databases and Ruby on Rails lately. I didn't get to meet the programmer beside him, but she had a copy of one of O'Reilly's Python books on her desk—good to see!

I realize medical devices aren't new, but underlying these medical devices are sophisticated models and a lot of computational crunching. It reinforced for me the revelation I had at Science Foo Camp in August—science has always had theoreticians and experimentalists, but the days of being primarily one or the other have passed in many fields. Increasingly we're seeing a new class of data crunching scientist, someone who can make sense of the enormous volumes of data that can now be gathered, someone who connects theory and practice with software. There's a lot of that at the Bioengineering Institute and it was very exciting to see up-close and personal.

Improving High School Science Education

As I read this fascinating NYTimes piece on a Florida teacher covering evolution, I was reminded of an interesting email exchange I had recently with Kevin Padian, a UC Berkeley professor in the Dept of Integrative Biology, and curator of the UC Museum of Paleontology. He was at Science Foo Camp, and afterward wrote in email:

My area is evolution, the most misunderstood concept in all of science. Two websites that help the public with this are at Berkeley's Museum of Paleontology (evolution.berkeley.edu) and the National Center for Science Education (www.ncseweb.org) (our ExecDir Genie Scott was one of the other participants at camp). I'm in the process of constructing a website on major transitions in evolution to which scientists can contribute, and which will be available to all teachers, students, and textbook writers. We really want to get this stuff into textbooks so that the creationist assertions that we have no evidence for microevolution can be countered. I've outlined a strategy for this in an article (PDF).

Kevin was kind enough to send me a copy of his paper. His thesis is that highschools react to college demands, so providing great free resources for college textbook authors will raise the bar for highschool textbooks. He points to a new type of illustration, the evogram (caution: long, see also this PDF of the relevant slide), which clearly shows evolutionary continuity over both organisms and time. He suggests evograms as a useful addition to the educational toolbox.

My reaction was that I didn't think targeting colleges would work:

I enjoyed your paper. I disagree with your pivotal assumption, though, that if colleges up their game then the high-schools will have to follow. That's just not the case--you only have to look to computer science to see how CS has been gutted at the high-school level. It's as though math were taught in high-schools as "how to use a calculator". Despite our jests, math isn't that bad in high schools--there's still serious math education happening even if it could be done better, but there's precious little serious computing education at all.

Kevin stuck to his guns, though:

Interesting observation, but I think you're making a slightly different (and highly valid) point: that "simplification" for lower grade levels can mean "dumbing down" or even "subvert crucial skills" (like using calculators for everything because they're not making kids learn multiplication tables, estimates, and so on). As a result, the whole structure of CS -- what kids need to know to be literate about CS at the HS level -- is lost. That's exactly what happens when the whole science of macroevolution becomes reduced to making "molds and casts" of fossils instead of teaching concepts about biodiversity through time.

But I will stick to my thesis: K-12 curricula won't include this stuff unless it's taught at the college level. Everything is downward-driven. High schools structure their course offerings based on what will get their kids into colleges. Even at the university level we structure our major requirements for many science departments based on what medical and professional schools want as preparation. It doesn't necessarily mean that if something is in a college text it will be taught in high school, but I'm making a different point: if it's not an important part of the college curriculum, it definitely won't be taught in K-12.

PS: take a look at our UCMP website, evolution.berkeley.edu to see what can be done informally to circumvent the usual textbook-curriculum-standards bottleneck.

I now agree with Kevin—if something's critical at college level, high schools will want to teach it and teach it well. I also love the idea of providing free educational materials that make it easy for textbooks and teachers to cover a topic well. It reminds me of csunplugged.org (created by Kiwis!) that Google funded to be publicly available at no cost. I'd love to see more organized efforts to improve the high-school and college education of science (and computer science) through small reusable teaching resources. Anyone know of some?

Radar Theme: Personal Genomics

[This is part of a series of posts that briefly describe the trends were currently tracking here at O'Reilly: 1, 2]

Genetic analysis software and hardware used to be very expensive, only for professionals—now it's trickling down to ProAms, and soon (under 5 years) will be widespread for consumer applications. This changes how drugs are developed and applied (don't test against 500 people and say whether it "works", figure out which genetic markers indicate the people it works for and sell to those), how diseases/conditions are diagnosed and treated, and our sense of self. Expect "interesting" (in the Chinese curse sense) interactions with privacy, workplace relations, and even parenthood.

Watchlist: 23andme, Hugh Reinhoff's "My Daughter's DNA".

Radar Theme: Neuro-everything

[This is part of a series of posts that briefly describe the trends that we're currently tracking here at O'Reilly]

Humans are consistently irrational, and every lottery ticket sold proves the point again. Psychologists, economists, neurobiologists are all studying what makes us behave the way we do. The promise is that we'll be able to be better: compensate for our biases, exploit the biases of others, or become smarter.

Watchlist: Mind Hacks, Nudge Blog, Robert Cialdini, Kluge.

Radar Theme: Synthetic Biology

[This is part of a series of posts that briefly describe the trends that we're currently tracking here at O'Reilly]

Drew Endy taught undergraduate students how to make e. coli bacteria that smelled like wintergreen, using his biobricks. This shows us a future for biology where "useful biological tasks" can be "automated" using "components". The quotes indicate where research and development are going—building components, figuring out how biological amateurs can assemble them, and to what end. The overlap with open source and the low-barrier-to-entry that's reminiscent of the web are particularly interesting to us.

Watch list: Drew Endy, George Church, Christina Smolke, Open Wetware, Ginkgo Bioworks.

Guessing gender from browser history

I just found a clever trick for guessing gender from browser history. I tried it and then realized that I'm a crappy test for the system: yes, likelihood of my being male is 99%. But if I read a hardcore geek tech blog, then that's probably the case anyway. I could emulate that behaviour with a simple return(G_MALE) in the code.

I pushed the link to a few women for some more strenuous testing. Penny Leach was told she's 52% likely to be female, and Laurel at O'Reilly was told she's 50% likely to be female. Perhaps on the internet, everyone surfs like a MALE with probability 50%. How'd the test work for you? Let me know in the comments ....

Should Personal Genomics Be Regulated?

I read recently about the cease and desist letters sent to 23andme and other personal genomics companies selling tests directly to consumers. 23andme has responded, saying that they agree with the ultimate need for regulation, but that harnessing the consumer internet for personal genomics is a really valuable scientific tool.

I have to say I find myself doubtful about the urgency of this regulatory move. It smacks more of the hand of the AMA, an entrenched industry trying to make sure that the new tools of genetic testing remain under the thumb of doctors, than of true consumer protection. You have only to walk into Whole Foods to encounter a multi-billion dollar industry of supplements making all kinds of dubious health claims, which is completely unregulated. Why pick on personalized medicine, which has way more substance, and at least so far, way more care in the types of claims it makes?

I'm a happy 23andme customer, and it's hard for me to see how the information they provide consists in any way of medical advice that should be subject to regulation. They are very clear to mark the scientific status of any genetic studies they report on, and never begin to presume to diagnostics. In many ways, the service is a kind of "RSS for the genome," feeding you the results of the latest scientific studies that might be relevant to you. You can find these same studies simply by googling for them. 23andme simply says "this might be relevant" using data that you've provided about yourself via your genetic sample.

The benefits of what companies like 23andme are doing is enormous. Once you understand even a tiny bit about how genes affect our response to drugs, you realize just how flawed many clinical studies are. If people have different genetically programmed response to drugs, what is the right dosage? Is the drug effective? (What if it's 100% effective for 3% of the population, due to their genetics, while it's completely ineffective for the rest. Would we ever know?)

Ideally, every clinical study going forward would have a genetic screening component. And new studies should be done on old treatments to correlate their effectiveness with genetic data. But that's hugely expensive.

23andme (and presumably other similar services) have come up with a very clever hack that will vastly increase the available store of genetic information that can be used to cross check various medical studies. They recruit a large population of those who are merely curious. But once that genotype data is available, they can begin to do surveys of their user population to gather corresponding phenotype data. (That is, matching genetic data with observable characteristics.)

This will be a boon to science. As Linda Avey, one of the founders, wrote on the 23andme blog:

Our first mission is to enable personal access to genetic information and provide a look, through the prism of an individual’s genome, at the flood of research discoveries being published. Our longer-term goal is to utilize a web-based platform that gives individuals the ability to share details related to their personal traits-including diseases they have and how they respond to therapies-uniformly layered on their genetic profiles to start building the evidence needed to drive targeted diagnoses and treatments.

I hope that regulators will seize the opportunity provided by the consumer internet to open up the frontier of personalized medicine. This is our data, the most personal data we have. I don't see why it should be forced into the straight-jacket of the 20th century medical industry. We tried that with music, and what did that get us?

I'd say to let this area run for a while, and start thinking about regulation if and when it starts to go awry. Right now, the companies I'm aware of are being very careful not to promise anything that needs to be regulated.