Direct To Consumer Genomics – A Long Way To Go 1


Image courtesy Libertas Academica via Flickr.

Once upon a time, when I was working for one of my prior employers (you can guess which one it was as the company name started with an ‘I’), the offer was made to have 23andMe to do whole-genome genotyping of myself and my immediate family. It being 2007 or so, the cost was not unreasonable in terms of the ’employee discount’ (23andMe used the company’s equipment and microarrays to do the genotyping) and after thinking about it for a few minutes, I decided to decline the offer.

Not many of my colleagues took the company up on the genotyping (if I remember correctly it was then priced in the few-hundred-dollar range, which was a nice discount at that time to the 23andMe commercial service), and the reasons why are understandable. When genomics becomes personal, I started thinking like an ordinary consumer, with ordinary consumer concerns.

These were a few of the concerns: what is the security of this information? Who has access to it? What if this startup company is sold to a large insurer, and even though GINA has been passed, what other unintended consequences may arise? And then of course, is the question of whether understanding my common variants is a worthwhile hobby. Genetics and genomics are topics that I concern myself with everyday in my ‘day job’, will I become a genomics hobbyist as well, delving into my own ancestry? (My brother and sister-in-law currently take up that hobby in my own family, and I don’t necessarily have a driving interest to supplement the work that they’ve already have done).

23 and Me

Years ago a geneticist friend told me that 23andMe masked out the most informative SNP on the microarray, due to the consequences of this information. It was for the ApoE gene, discovered by Margaret Pericak-Vance when she was a graduate student in Alan Roses laboratory at Duke. It is a famous paper, one of the most often-cited papers in biomedical research, and way back in 1993 she did it via linkage analysis. The APOE4 variant is the largest known genetic risk factor for late-onset sporadic Alzheimer’s Disease (AD), and if I am going to get my genetic material genotyped I would like to have that information. (I would note that I do not know whether they return the APOE result today, and if you do happen to know please provide a comment.)

Which leads to the natural concern – what to do with this massive amount of variant data, and what does it mean? And this doesn’t begin to touch the information available if I paid a bit more to get my own exome sequenced, which 23andMe offered on a trial basis (now closed). By ‘massive amount’, the Infinium HumanHap550 way back in 2007 produced over 550,000 single nucleotide polymorphisms (SNPs) per individual, with a population minor allele frequency of 5% of greater (it being based at that time on variation data from the HapMap project), whereas now there are over 4M SNPs on an Illumina Omni5-Quad with an MAF of >1%. From the 23andMe.com website, they now are delivering over 1M SNPs for $299.

Is a genotype useful?

Millions of variants, and 1,355 Genome Wide Association Studies (GWAS) published as of the summer of 2012 you can dig into the databases that have been setup by the NHGRI and others to display odds ratios and p-values for particular variants, and end up actually where you began. A NY Times article from an early user of 23andMe’s service in 2007 had the conclusion, that the current state of interpretation meant that all the information they received from the report was indicated already by their family history, and the additional information was more for entertainment purposes. (Items such as the fast twitch / slow twitch myosin fiber ratio, or the dryness of your earwax.)

Everyone who has had a physical exam understands the importance of family history information as it impacts your own health; susceptibility to heart disease, whether cancer ‘runs in your family’, etc. And this information is invaluable in guiding healthcare, whether for more frequent screening of a particular biomarker or a specific recommendation for lifestyle changes.

23 and Me offers consumer genomics, and other companies have been founded on that same premise such as Navigenics (recently acquired by Life Technologies) and Pathway Genomics. Several other small companies do single-gene genotyping on a custom genotyping panel, using equipment from Sequenom or even Sanger sequencing. In 2010 the General Accountability Office (GAO) issued a report to Congress indicating the results of an undercover investigation where identical samples were submitted to 15 direct to consumer (DTC) companies offering genetic testing, with conflicting results and erroneous claims.

From the GAO report: “Posing as consumers seeking information about genetic testing on the Internet and through phone calls and face-to-face meetings, we found that 10 of the 15 companies we investigated engaged in some form of fraudulent, deceptive, or otherwise questionable marketing practices.” And for the regulation of this new industry, the FDA is looking at how to properly regulate it, from a medical device / medical diagnostic testing point of view.

23 and Me offers consumer genomics, and other companies have been founded on that same premise such as Navigenics (recently acquired by Life Technologies) and Pathway Genomics. Several other small companies do single-gene genotyping on a custom genotyping panel, using equipment from Sequenom or even Sanger sequencing. In 2010 the General Accountability Office (GAO) issued a report to Congress indicating the results of an undercover investigation where identical samples were submitted to over 10 direct to consumer (DTC) companies offering genetic testing, with conflicting results and erroneous claims.

From the GAO report: “Posing as consumers seeking information about genetic testing on the Internet and through phone calls and face-to-face meetings, we found that 10 of the 15 companies we investigated engaged in some form of fraudulent, deceptive, or otherwise questionable marketing practices.” And for the regulation of this new industry, the FDA is looking at how to properly regulate it, from a medical device / medical diagnostic testing point of view.

In the case of 23 and Me and Navigenics, these DTC genomics companies hired genetic counselors to aid in the interpretation of the results, enabled the ability to look at ancestry, and in general broke new ground in consumer awareness. But given the statistical genetics involved, and risk factors associated with particular variants, with 500K to several million SNPs to look at for any given individual this task is non-trivial, to rank variants on its relative basis for risk and health impact to the individual. Thus the need for experienced counseling, to accurately portray how a given individual should behave with an avalanche of information.

DTC Exome sequencing

Now when an individual exome is sequenced, there are many more variants – tens of thousands – with varying degrees of  impact on individual genes and their translated proteins. Earlier this year in May 2012 the Broad and Baylor genome centers published a remarkable paper in Science reporting the results of no less than 2,440 human exomes at an average depth of 111x. Their findings? >500,000 variants discovered, the majority of them (86%) rare (<0.5% allele frequency).

On an individual level, the average person in this study had 13,595 Single Nucleotide Variants (SNVs), and of those ~313 of these SNVs affected protein function. (The paper used no less than seven different measures of non-synonymous coding variants’ impact on functional protein, and took a ‘majority rule’ approach of four of the seven methods giving a positive signal.) Having hundreds of genes that have damaging mutations in them may seem surprising, that somehow we can compensate for the lack of so many functional genes. And the paper goes into the implications of this ‘explosion’ of variants in recent human evolution.

And then what of interpretation of all these individual gene functions? For a given individual who is considered in good health, what kind of interpretation could be offered?

The PGP

Another recent paper, published by the Personal Genomes Project (http://www.personalgenomes.org), sheds some insight into how this data could be used. Taking the first 10 individuals who have fully consented as part of this project to not only have their genomes sequenced and analyzed, but made public including their medical history and other phenotype information, these individuals are genomic pioneers. (The PGP has already enrolled some 1,800 individuals.)

An analysis pipeline was developed by the PGP, called the Genome-Environment-Trait Evidence (GET-Evidence) system, which prioritizes variants that are both published and novel.

Looking at the first 10 samples of putatively healthy individuals, there are “numerous literature references implying serious disease”, which due to the nature of this study was likened to an ‘incidental finding’ (a term borrowed from the medical discipline of Radiology, but others have made a case that instead of the term ‘incidental finding’ the term ‘variant of unknown significance’ be used instead). In one example of a highly penetrant variant for a late-onset disease, on individual (identified as PGP6) has a variant MYL2-A13T which causes familial hypertrophic cardiomyopathy, a potentially lethal heart condition. The paper has a number of other instances of moderate-severity variants, and how each was handled from a counselling and how this information is communicated.

So there’s a long road ahead for consumers to understand what their genetic makeup actually means, and in the meantime genomic research continues apace.

References:

Tennessen and Akers et al., Science May 2012, “Evolution and Functional Impact of Rare Coding Variation from Deep Sequencing of Human Exomes”.

Ball and Church et al., Proc. Nat. Acad. Sci. July 2012, “A public resource facilitating clinical use of genomes”.

For a personal computational biologists’ analysis of his own exome, Jung Choi from Georgia Tech in Atlanta has written up a few posts about it, to give you an idea of how complex it is.

 


About Dale Yuzuki

A sales and marketing professional in the life sciences research-tools area, Dale currently is employed by Olink as their Americas Field Marketing Director. https://olink.com For additional biographical information, please see my LinkedIn profile here: http://www.linkedin.com/in/daleyuzuki and also find me on Twitter @DaleYuzuki.

Leave a comment

Your email address will not be published. Required fields are marked *

One thought on “Direct To Consumer Genomics – A Long Way To Go