Tag Archive for '23andme'

Page 2 of 2

Finding the holes in our genomes

In a previous post I discussed copy number variation, a form of genetic variation not broadly reported by DTC companies. In today’s post I provide a very simple program that allows one to identify potential deletions on the basis of high density SNP genotypes from a parent-offspring trio, and report on the results of running this program on data from my own family.

The program uses an approach that I applied as a graduate student to mine deletions from the very first release of data from the International HapMap Project in 2004.  The idea, explained in my last post, is to look for stretches of homozygous genotypes interspersed with mendelian errors, which might indicate the transmission of a large deletion. Let’s be clear, this is a simple analysis that most programmers and computational biologists would find straightforward to implement. It is probably a good practice problem for graduate students and would-be DIY personal genomicists.

I obtained 23andMe data from both my mom and dad, and, with their consent, ran the three of us through the program. I was mildly surprised to find only two potential deletions; I had previously speculated that one would find 5-10 deletions per trio with the 550K platform used by 23andMe.

Continue reading ‘Finding the holes in our genomes’

Am I partly Jewish? Testing ancestry hypotheses with 23andMe data

I agreed to make my 23andMe genotyping results publicly available as part of GNZ without a moment’s hesitation. This is in part because I knew the results were actually a bit dull (in a good way, I suppose) – I’m not at vastly increased or decreased risk for any diseases (based on research so far), and I was unsurprised to find out that I have blue eyes. I was also unsurprised that 23andMe identified me as most likely of north European ancestry.

Several hours after we released our data, however, I was pointed to a post where Dienekes Pontikos wrote about the results of running all our data through his ancestry prediction program. While just about everyone was quite confidently predicted to be almost entirely of northwestern European descent, this analysis gave me a point estimate of 20% Ashkenazi Jewish ancestry. Within hours, several people had asked me about this, and I had no real response. So I decided to take a look at the data myself; some basic analyses are below.
Continue reading ‘Am I partly Jewish? Testing ancestry hypotheses with 23andMe data’

Getting even with the odds ratio

In the recent report from the US Government Accountability Office on direct-to-consumer genetic tests, much was made of the fact that risk predictions from DTC genetic tests may not be applicable to individuals from all ethnic groups. This observation was not new to the report – it has been commented on by numerous critics ever since the inception of the personal genomics industry.

So, why does risk prediction accuracy vary between individuals and what can be done to combat this? Are the DTC companies really to blame?

To explore these questions it is first necessary to understand what is meant by the odds ratio (OR). In genetic case-control association studies the OR typically represents the ratio of the odds of disease if allele A is carried compared to if allele B is carried. If all else is equal, genetic loci with a higher OR are more informative for disease prediction – so getting an accurate estimate is extremely important if prediction underpins your business model. However, getting an accurate estimate of OR is far from easy because many, often unmeasured, factors can cause OR estimates to vary. In this post I will try to break down the concept of a single, fixed odds ratio for a disease association, and highlight a number of factors that can cause odds ratios to vary using examples from the scientific literature.

Continue reading ‘Getting even with the odds ratio’

Friday Links

Over at 23andMe’s blog The Spittoon, Stanford genetics professor Uta Francke has a great point-by-point dissection of the new policy statement of the European Society of Human Genetics on direct-to-consumer genetic testing. Francke doesn’t shirk from explaining that this statement should be seen in the context of an ongoing turf war between traditional geneticists and DTC upstarts:

For example, organizations like ESHG, the American Society of Human Genetics (ASHG), and the American College of Medical Genetics (ACMG) were created to represent the interests of their professional membership, similar to the American Medical Associations (AMA) for physicians. Any claims to act in the public interest by protecting people from potentially damaging genetic information may reflect a fear of the new consumer-driven healthcare system that, as in the case of DTC genetic testing, may circumvent the professional establishment. Insisting on individual professional counseling in the pre-testing and post-testing phase can be interpreted as an attempt to ensure continued involvement of board-certified genetics professionals. [DM]

In PLoS Genetics, researchers report that a SNP associated with risk for colorectal cancer likely exerts its effect by modifying the expression level of a nearby gene. Previous studies on a different region associated with both prostate and colorectal cancer revealed a similar mechanism (see here, for example). This contributes to growing molecular evidence that SNPs with long-range regulatory effects might be generally important in disease. [JP]

Kevin Davies’ superb new book The $1000 Genome is now available on Amazon. Keith Robison has already posted his review, and I’ll have my full thoughts here on Genomes Unzipped soon – but for now I’ll just say that Davies’ long and extensive experience working the genomics beat really pays off, with both a wealth of inside information and juicy anecdotes about the key players in the field. Well worth a read for anyone interested in modern genomics (i.e. everyone reading this post). [DM]

Dan Koboldt of MassGenomics has a useful review of the recent Cold Spring Harbor Personal Genomes meeting. He pulls out four key themes emerging from the conference: new estimates of human mutation rates, more sequencing of cancer genomes, studies of genome regulation and epigenetics, and an explosion of exome sequencing in both severe and common diseases. [DM]

Our own Dan Vorhaus reports on the utterly bizarre decision by Health and Human Services (HHS) Secretary Kathleen Sebelius and NIH Director Francis Collins to discontinue the Secretary’s Advisory Committee on Genetics, Health, & Society (SACGHS). Inexplicably, Sebelius and Collins argued that “the major topics related to genetic and genomic technologies had been successfully addressed by the committee through its comprehensive reports and recommendations over the years”. Vorhaus notes diplomatically that “it is clear that even those issues SACGHS investigated in detail have not been resolved with any meaningful degree of finality”. [DM]

Estimating the size of the DTC genomics market

Over the last few years, I’ve found that the same question keeps cropping up again and again at meetings whenever we talk about direct-to-consumer genetic tests:  “How many people are actually buying these tests?”. And because the companies (for whatever reason) have thus far been rather reticent about telling us how many kits they’ve sold, until recently the answer has simply been “I don’t know”. Yet if we’re going to talk about their sociological impact, their knock-on effect on health systems, and re-writing our regulatory laws around them, surely this is something we ought to have a handle on.

So… how many people have actually bought these tests then?

The problem is how to go about estimating a market size when there is precious little data, and the companies are all privately owned? First, we teamed up with some enthusiastic MBA students, who came up with the simple but elegant idea of using website hits as a proxy for market share. Using Compete.com, we found that the ‘big three’ – 23andMe, deCODEme and Navigenics – together had just over 662,000 unique hits during 2009, of which 23andMe received the lion’s share at nearly 80%. (They received fairly constant internet traffic throughout the year, with an average of around 43, 4 and 8 thousand unique visitors per month respectively). Pathway had only just launched when we did the analysis, resulting initially in a large transient spike in internet traffic, so we left it out.

Second, fortunately for us, in October 2009 23andMe stated publicly that their database contained “30,000 active genomes”, which were either sold or given away at a substantially reduced rate. (This rose to 50,000 in June 2010, but that doesn’t really alter the calculations). So, assuming a steady rate of uptake, this equates to perhaps 15,000 genome scans sold during 2009. Combining this with the internet traffic data, we estimate in this month’s Genetics in Medicine [Wright CF, Gregory-Jones S. Genet Med. (2010) 12: 594] that around 20-30,000 genome scans were sold in 2009, at a cost of between $300-1,000, which probably equates to a commercial value of around $10-20 million.

Is that really true?

Obviously there are substantial margins of error in any estimate made from such limited data, and caveats include the fact that we only considered tests sold during 2009 and we ignored (as much as possible) non-medical tests like paternity and ancestry testing. Nonetheless, this seems like a realistic ballpark figure, and importantly it is neither millions of people, nor hundreds of millions of dollars. We don’t know (yet) how big the market for whole genome sequences will be, or what impact preconception carrier testing might have, but at the moment it is clear that the market for DTC genetic testing is much smaller than expected or than one might surmise from all the media attention. Which means that the alleged harms to consumers, and the reputed knock-on effects on health systems, must necessarily be limited.

Nonetheless, I would dearly love to hear from any DTC genomics companies out there willing to share some more concrete data…

Communicating genetic data to DNA donors

The work of geneticists, a category that includes  the majority of Genomes Unzipped contributors, typically consists of analyzing DNA sequences from large collection of individuals and this constant flow of data gives us an overview of the diversity of human genotypes. And while in most cases these mutations do not have any functional impact, some rare cases are well documented and have important adverse effects.

A famous example is the BRCA2 gene for which rare mutations have been linked to an increase prevalence of breast and ovarian cancer. Another example: multiple rare variants have been linked to various forms of familial hypercholesterolemia, a condition that significantly increases heart disease risk. I picked these examples because for both cases the identification of carriers of these rare mutations in the general population could improve health: aggressive detection of breast cancer, and use of relevant treatments (such as statins) if you are a familial hypercholesterolemia patient, can make a real difference.

The fact that, in some cases at least, something can be done can put geneticists in a difficult situation. Indeed, we often come across known disease related mutations in the DNA from patients who were not recruited for anything linked to that disease. And it is not clear how this information should be handled. On one hand, we cannot assume that the patient has any desire of knowing anything about his/her disease risk. On the other hand, while analysts always work on anonymous genetic data, the medical staff that collected the sample could potentially get back in touch with the patient who donated his/her DNA. Letting DNA donors know may actually make a difference in their lives (again, this situation is rare but it happens).
Continue reading ‘Communicating genetic data to DNA donors’


Page optimized by WP Minify WordPress Plugin