The 23andMe Personal Genome Service™ offers information about customers’ maternal and paternal ancestry by examining their mitochondrial DNA (which we all inherit from our mothers) and the Y chromosome (which is passed by fathers to their sons).

Over our species’ history new genetic variations have arisen spontaneously in those pieces of DNA, and been passed down through the generations. So researchers know that every person alive today who has a particular variation — say a T instead of a G at a certain spot on the mtDNA — is descended from a single common ancestor. By sampling people from around the world, scientists have been able to assemble “family trees” that trace all the way back to the dawn of the human species in eastern Africa more than 100,000 years ago. Those trees are the basis of the haplogroup assignments we give our customers. Each haplogroup represents a particular branch — and therefore a unique sequence of SNPs.

Now you can see exactly which SNPs we use to generate the 23andMe mitochondrial and Y chromosome haplogroup trees, using a feature we’ve developed called the Haplogroup Tree Mutation Mapper. This isn’t the kind of information everyone would necessarily want, but a number of customers who are especially interested in genetic genealogy have requested it. It can be used to compare our haplogroup assignments to those obtained elsewhere, for example.

The Haplogroup Tree Mutation Mapper is the first arrival in 23andMe Labs, our new technology sandbox where we will showcase experimental features not currently available in our Personal Genome Service™.  These features may still be in development, require specialized knowledge or be of interest only to a subset of our customers. Each lab will have its own community so customers can compare notes, ask questions and share ideas.

Some labs will be requested by customers. In fact, we welcome your suggestions. Others will be dreamed up by our scientists.

Finally, you can expect labs to be a little less refined than what’s available within the Personal Genome Service, and somewhat fluid as well. A feature could be discontinued at any time, or it might be elevated to full integration with our Personal Genome Service.

We’re really excited about having this new outlet for our ingenuity, and we hope it will engage some of yours as well!

ShareThis

Editor’s Note: This week TIME Magazine is naming the 23andMe Personal Genome Service™ its Invention of the Year, an honor that the publication has previously bestowed on innovations such as the iPhone and YouTube. This post by Director of Products Alex Wong (back row, second from right) offers a glimpse at how 23andMe came to be, and what makes it such a powerful tool for exploring your DNA.

We’re thrilled that our product has been named TIME Magazine’s 2008 Invention of the Year. But what exactly did we invent? To answer that question, let’s to go back to the early days of 23andMe … .

I first met our co-founders Anne and Linda, together with Serge and Brian, our first two employees, in the summer of 2006 at a bakery in downtown Palo Alto, Calif.. As they were explaining the basic concept behind the company, Linda reached into her bag and pulled out a CD with “Aveys” written on it in magic marker.

“What’s that?”, I asked.

Her reply: “It’s my family’s genomes!”

That’s a first, I thought to myself — pulling genomes from a handbag. I was hooked. I knew I wanted to join 23andMe and help make this possible for everyone and their family.

But we had a long way to go. DNA chip technology was still fairly new, something that only researchers had access to. Linda, having worked at a DNA chip company, was able to get her own family genotyped. But unless you had an in at a research lab or one of these companies, getting your own genome just was not possible. Even if you did manage to get the genomes of you and yours on a CD, all you’d find when you looked at it would be a bunch of giant, indecipherable text files filled with “rs” numbers and A’s and B’s. These are the raw data files of SNP information that the DNA chip analysis produces.

How to make sense of all that data? The good news was that years of research, starting with the Human Genome Project and the International HapMap Project, had produced a wealth of publicly available information about human genetics. The bad news was, it was all locked inside databases designed by researchers, for researchers. Want to know what one of your SNPs means? A logical place to start, you would think, would be dbSNP, a massive database of SNPs run by the National Institutes of Health.

Good luck. Assuming you can wade through its dense user interface, make sure you get your alignment, genome build, major and minor alleles, orientation, and stranding right, or you’re liable to get your interpretation backwards and think you’re supposed to have wet earwax when you really have dry! Don’t know what stranding, genome builds, or any of those other things are? Trust me, you’d rather not.

Not only did these government and academic databases sport slow and unfriendly user interfaces, but some of them also had data quality and consistency problems and few of them were linked together.

So it was going to require more than just building a website to make all the fruits of ongoing genetic research accessible, useful, and compelling to regular people like you and me. We had to create a whole architecture and process for cleaning up, aggregating and integrating all of this human genetic information, combining it with an individual’s own genotype data, running comparisons and algorithms on it, blending it with great educational content, and serving it back to our customers through a friendly user interface — all in under 500 milliseconds, which is how long you have before people start to complain that your website is slow.

On top of that, we wanted personal genetics to be social, so our customers could see their genetic data in context and learn about what makes us all different and similar at the same time. That meant building in the ability for customers to share their data with family or friends without sacrificing privacy or security.

It’s taken the combined talents of all the different kinds of people we have at our company, from bioinformaticians, to engineers, to writers, to UI designers, but we’re all so excited about having made it possible in 2008 for anyone with $399 to explore their own genome.

Of course, if you’d like to rock it 2006-style, you can still get your own gigantic text file full of SNPs using our Download Your Raw Data feature. You can even burn it onto a CD and stick it in your handbag!

ShareThis

This is what the new view looks like:

The Clinical Reports page (above) shows topics containing information that is both supported by solid scientific evidence and has the potential to make a real difference in terms of your disease risk or odds of having a particular trait. There are four different sections within Clinical Reports:

• Disease Risks shows you the currently covered health conditions for which your risk is most elevated.
• Carrier Status tells you whether you have one or more of the disease-causing mutations we cover.
• Traits tells you what your genes indicate with regard to a number of physical attributes, from earwax consistency to malaria resistance.
• New and Recently Updated is just that — a list of our newest and most recently updated Health and Traits topics.

The second page of the dashboard contains a list of Research Reports; these contain information that either has not yet been scientifically validated or has been, but does not have a substantial influence on your odds of having a disease or trait. On the left side of the list there are colored dots that indicate what effect, if any, your data have on your genetic risk of disease.

Here’s what the Research Reports page looks like:

The new Health and Traits dashboard isn’t intended to give you all of your information on a single page — you still need to look at the individual reports to see what they contain. But as the number of diseases, conditions and traits we cover continues to grow, we hope it will help our customers find the ones that are of greatest interest to them!

ShareThis

The judges have met and a winner has been chosen.

In the first 23andMe Win Your Genome Contest, the challenge was to describe Lilly Mendel – a real person whose data are presented in the 23andMe demo account – based on her genetic information alone. As we declared in the announcement of the contest, entries were judged based on accuracy, creativity and cleverness.

Cleverness we got. More than one entrant actually claimed to BE Lilly Mendel. Many entries employed flattery by concentrating on Lilly’s physical attractiveness, intelligence and athletic ability – very clever.

As for creativity, we have no idea where some of you got the idea that Lilly’s family harbors a secret fear of “gigantism,” that she enjoys dancing the tango or that her nose twitches when she gets angry. And though her genes may indicate a preference for the bignay fruit, we have no idea if she has ever tasted it.

But the clear winner in all three categories was Mike Cariaso, who sent us a detailed description based on data from the SNPedia database, an online compendium of genetic associations. Using software he has written (and made publicly available) that evaluates raw 23andMe data using the SNPedia database, he assembled a formidable description of Lilly. Its accuracy may have suffered a bit from the author’s speculation, but the creativity of those conjectures more than made up for it.

Click here to see Mike’s entry, which he posted on the SNPedia page (you can also see it after the jump). The description is annotated with SNPs and brief descriptions of what he gleaned from each of them.

As the winner of our first contest, Mike receives the 23andMe Personal Genome Service™ for a person of his choice. We thank everyone else who entered, and encourage you to enter the next one, which should be announced in a few weeks.

Lilly Mendel is a female caucasian

• rs1426654(A;A) probably light-skinned, European ancestry
• rs2814778(T;T) Duffy blood group antigen. suggests european ancestry
• rs25487(A;G) 2x higher risk for skin cancer

with beautiful blue eyes,

• rs4778241(C;C) blue eyes
• rs12913832(G;G) blue eye color
• rs7495174(A;A) blue/gray eye color
• rs8028689(T;T) blue eye color if part of blue eye color haplotype
• rs7183877(C;C) blue eye color if part of blue eye color haplotype
• rs1800401(G;G) blue/gray eyes possible

brunette hair

• rs1805005(G;G) not a blonde
• rs1805008(C;C) not a redhead
• rs2228479(G;G) not a blonde, nor a redhead

and an interesting personality.

• rs4570625(G;G) higher scores on anxiety-related personality traits
• rs1800955(C;T) influences personality

Everyone’s who’s ever met her knows that. Fewer people know she’s a night owl

• rs1801260(C;T) late night / light sleeper

who likes to finish her evening with french vanilla ice cream and a glass of white wine.

• rs4988235(T;T) can digest milk
• rs10246939(C;T) can taste bitter
• rs1726866(C;T) can taste bitter

Even fewer people know of the signs of alcoholism in her family tree

• rs279871(C;T) rs279845(A;T) rs279836(A;T) =~ haplotype rs279871(A) rs279845(T) rs279836(A)
• (this needs orientation confirmation)

Or how those family issues may have helped to trigger a period of depression during her teens. After experimenting with several medications doctors found that citalopram seemed to work well for her.

• rs2156921(G;G) increased risk of depression
• rs2235015(G;G) 7x less likely to respond to certain antidepressants
• rs2032583(T;T) 7x less likely to respond to certain antidepressants
• rs2235040(G;G) 7x more likely to respond to certain antidepressants
• rs2235067(G;G) 7x more likely to respond to certain antidepressants
• rs12720067(G;G) 7x more likely to respond to certain antidepressants
• rs1954787(C;C) depression is ~10% more likely to respond to citalopram
• rs2518224(A;A) citalopram will not increase suicidal thoughts

Earlier genetic testing might have prevented 8 months of frustrating experimentation.

In her mid 30s she experienced a blood clot. By now genome scans were a part of a patient’s electronic medical records. Her tests revealed the need to prescribe a smaller than usual dose of warfarin

• rs8050894(C;C) warfarin sensitivity

This was the excuse she’d needed to quit smoking. Doing so was surprisingly easy since there were now medicines customized to her genetics.

• rs1800497(C;C) A2/A2 bupropion effective to quit smoking

That same report revealed she had dyslexia

• rs761100(C;C) higher risk for dyslexia
• rs3212236(T;T) higher risk for dyslexia
• Based on rs793862(A;G) rs807701(G;G) genotypes she must be carrying one copy of the haplotype rs793862(A)-rs807701(C) which is linked to increased risk of dyslexia.

which she’d easily overcome

• rs363039(C;C) increased performance IQ
• rs363050(A;G) 3+ IQ points
• rs17070145(C;T) increased memory performance
• rs760761(C;T) and rs2619522(G;T) heterozygotes for cognitive function

She hadn’t expected to see that, but somehow she felt she’d always sort of known. What came as a surprise was that instead of the usual advice to exercise more she learned that her jogging might actually be something to watch out for

• rs1815739(T;T) endurance athlete
• rs1024611(C;T) increased risk of exercise induced ischemia

Based on her specific genetics

• rs10494366(T;T) Shorter QT interval
• rs9739493(T;T) higher risk for hypertension
• rs173686(G;G) stroke risk
• rs326(A;A) lower HDL cholesterol
• rs2412522(G;G) Framingham related to cardiovascular disease for mean corpuscular hemoglobin concentration

Her doctors opted for a daily statin to help keep her heart healthy

• rs1799768(-;-) a deletion which increases risk of myocardial infarction
• rs3900940(C;T) increased risk of coronary heart disease; better response to statins
• rs2298566(C;C) increased risk of coronary heart disease; better response to statins
• rs5918(C;T) heart attack risk, aspirin resistance

She’s become more vigilant about regular mammograms.

• rs1008805(C;T) increased risk of breast cancer
• rs931127(A;G) 2.15x more likely to have lymph node metastasis
• rs351855(C;C) inefficacy of herceptin

And learned that as she ages those beautiful blue eyes may fail her

• rs2165241(C;T) 4.4x increased risk of exfoliation glaucoma
• rs1329428(G;G) 2x increased risk for macular degeneration

So she’s started some daily eye exercises, become a participant in a research study and taken up photography.

ShareThis

Some customers who recently ordered our Personal Genome Service™ have had to wait more than the usual four to six weeks after spitting to get their genetic information. We know that it can be frustrating to wait when you are curious about what your genome holds, so we would like to apologize and offer an explanation for the delay.

Because 23andMe is creating an entirely new kind of business in delivering personal genetic information, the regulatory requirements we face are both complicated and uncertain. Because of the way these requirements are evolving, we recently changed the laboratory where our customers’ saliva samples are processed.

Anne Wojcicki

Our new contract laboratory is certified as a “high complexity” laboratory under the federal Clinical Laboratories Improvement Act of 1988 (CLIA). Though CLIA certification and oversight are appropriate for specific health and disease-related testing, we are complying with these guidelines to be consistent with other types of laboratory testing.

This switch does not affect in any way the nature or quality of the information we provide. We are still using the same technology as before. However, the transition from one lab to another has resulted in delays that are temporarily increasing the length of time it takes to process customer samples.

At no time have we experienced any problems with data quality. We are confident, after comparing genome scans from the previous and new laboratories, that they produce consistent information at greater than 99.9 percent accuracy. That means customers who had their samples processed before the switch will not be at a disadvantage now or in the future.

Again, we regret the inconvenience to our new customers. We expect to be able to resume our normal pace of operations within a few weeks. And in the long run, we believe this switch will benefit all of our customers. We’re excited to continue building the 23andMe community!

ShareThis