For some people, genetic information is formidably powerful. It can reveal that you have inherited a debilitating disease which lies unavoidably in your future, that you have a massively increased susceptibility to a particular cancer which can only be mitigated by surgery, or that you are not biologically related to your parents and siblings.
But, for many people, it’s actually quite mundane and uninformative. I’m one of those people. Undergoing genome-wide profiling was interesting, educational and worthwhile, and I would certainly recommend it as a voyage of exploration. But it hasn’t really been useful for my health. (It’s perhaps worth noting that a SNP profile covers only a fraction of the potential disease-causing variants, but nonetheless I doubt there’s anything I need to be scared of lurking in the rest of my genome.)
So, what did I learn?
From both 23andMe and deCODEme, I’ve discovered that, regardless of my own personal genetic risk, I’m at a distressingly high risk of various common diseases and a vanishingly small risk of numerous rare ones. This is certainly interesting from an epidemiological perspective, and one could argue that learning more about the prevalence of different diseases is one of the major educational benefits of personal genomics. But although I learnt a bit about the many diseases out there (and have marvelled at the fact that any of us are ever healthy), I haven’t exactly been moved to do anything as a result of my genetic information.
I’m at a higher risk of both types of diabetes relative to the general population, a fact which certainly could be useful if I (a) didn’t already know it from a distant family history of both diseases, (b) didn’t already maintain a healthy weight by eating well and exercising (fairly) regularly and (c) was really, honestly going to change my behaviour as a result of the information. Which, as we all know, is easier said than done. Intriguingly, my highest genetic risk is actually for rheumatoid arthritis (RA), but a little probing into the details indicates that this is primarily due to being homozygous for minor alleles of two closely linked SNPs (rs2476601 and rs667967) in the PTPN22 gene on chromosome 1… which is exactly the same gene that apparently underlies most of my genetic risk to type I diabetes. This left me wondering if I really am at higher risk of RA, or – given my family history – if in the context of my personal genetic background, this risk factor is more likely to manifest itself as diabetes. And this, of course, is part of the problem of trying to apply susceptibility variants uncovered in population studies to individuals.
On a brighter note…
It turns out that I’m at a substantially lower risk of age-related macular degeneration (AMD) than most people, which is certainly good news since this is quite strongly heritable, but not something I was losing sleep over to start with. (I’m well past the age to worry about type I diabetes, but well shy of the age where I might consider worrying about AMD!)
I don’t really seem to carry any of the common mutations tested for recessive Mendelian diseases, which lacking a family history in any of them is not exactly a surprise (though is still not definitive). Learning one’s carrier status is certainly something that seems potentially useful – particularly in conjunction with a partner prior to considering having a child – but for someone with no immediate plans to have children, the information is of rather limited relevance. Similarly, not currently being on any drugs, the pharmacogenetic information is little more than a curiosity to me, and I worked out years ago that I must be a fast caffeine metaboliser. (I am also unsurprisingly European by descent, but am pleased to discover that I might have blue eyes after all, contrary to the rather more persuasive phenotypic data…)
Don’t be surprised or disappointed by a ‘boring’ genome
Jokes aside, I have previously noted that many people feel rather under-whelmed by the experience of having a genome scan. But I would suggest that it’s the wrong message to take from the experience, and here’s why: if you go to your GP and ask for a health check, most people would be quite pleased to be told there was nothing wrong. I’ve taken home glucose and cholesterol tests, both with ‘normal’ results… this is good news, surely? So why be disappointed with an (apparently) uninteresting genome? If anything, I should be pleased. Of course, there is no such thing as being genetically normal – and I’m as much a mutant as the next person – but we shouldn’t expect to find hidden treasures in everyone’s genomes, nor skeletons in the cupboard. It’s just one of many molecules involved in the mind-bogglingly complex biological system that makes me me and you you. It might be useful in the future, and it might not. But either way it’s definitely interesting, and although that’s not a good enough reason for a health service to offer genome sequencing to everyone, it is a good enough reason to allow people access to a little personal genomic tourism if they want.
As technology improves and science progresses, we will understand an increasing amount about the genome; it will become a standard part of medical practice in many clinical contexts, and ever more individuals will have the opportunity to explore their own genetic information. But we shouldn’t assume that it will be useful for all – some people could benefit enormously right now from this kind of personalised genomics, and some people will never benefit directly. For now, what’s important is that we approach these tests with a realistic sense of what to expect, armed with good information, an open mind, and a healthy dose of common sense.
RSS
Email
Twitter
Authors Elsewhere
Refreshing perspective.
The most interesting things to come from my 23andMe SNP screen:
Running Dienkes’ ancestry predictor and having the results match closely with my (hybrid, European) family history.
Looking up ApoE genotype (strong Alzheimer predictor), not finding it, then learning about Bob Greene’s studies on how people deal with the prospect of carrying a high risk for a terrible chronic disease (mostly, extremely well).
Corresponding with people who 23andMe’s algorithm suggest might be third cousins, or so. No confirmed hits yet…
Looking up SNPs as they appear in the literature — it’s surprising how many are in 23’s database.
I cannot find if it was something like 23andme or whole genome sequence? Not sure if something like 23andme can really be called ‘genome’ !
sm,
This was both 23andMe and deCODEme SNP chip data. In the sense that these data provide a (limited) winbdow into Caroline’s genome, I think her terminology is appropriate here. Of course, she (and the rest of us) may well discover far more interesting things as and when our complete genome sequences become available…
Sorry, yes just SNP data – I added a clarification of this point into the text.
“I’m at a distressingly high risk of various common diseases and a vanishingly small risk of numerous rare ones. This is certainly interesting”
Not really. It is exactly what you would expect from probability.
It seems that these SNP panels aren’t revealing much more information than we can get from our family history, which makes the arguments against genome scans less forceful. Doctors have relied on indirect genetic information for decades.
Except for rare instances of large effect alleles, the best predictors of your future health course are still your personal health history and your family history. There’s a trade off at work: as genomics improves and uncovers smaller effect alleles, they simply become less important. It’s great as an academic exercise that we can uncover SNPs that account for 5% of the variance in a trait / disease, but what practical relevance does that have to most people?
“So why be disappointed with an (apparently) uninteresting genome? If anything, I should be pleased.”
For one thing, it still costs a lot for most people. You can compare it to car insurance. I hate writing that huge check every six months, and if it wasn’t legally required, a lot of people wouldn’t do it. It’s only in the rare instances when disaster strikes that, ironically, it feels like paying car insurance was worth it.
congratulations on being boring! on the disease front at least.
Not really. It is exactly what you would expect from probability.
i assumed she was joking dawg ;-) did i misinterpret?
23andMe is expensive but achievable. deCODEme is definitely beyond most people’s budgets, certainly mine.
There are a few problems with the SNP scans using that chip technology. One is that there isn’t a lot of information on the SNPs and what they do or don’t do. SNPedia has information but often times it just lists the SNP and mentions what it supposed to do but with no risk estimates or even which allele combination is better, neutral or worse. So, the lack of information is a problem.
Another problem is that 23andMe, and deCODEme offer a general product, your Raw Data, with little use for it. I have no interest in disease predictions, genealogy or finding cousins I was blithely unaware. Once you have your Raw Data, what is the use of it? The second problem is lack of applications to use this Raw Data. I have given my Raw Data to Polako, to Dienekes, to Promethease Reports and some others because I am trying to get more usage out of this data. What have I learned? Nothing much. Apparently, I come from the place I was born in, and where my ancestors come from. I knew that, as I am mono ethnic for hundreds of years. I have found out I share bits of dna with total strangers via 23andMe, HIR Search and Polako. I have found that I have minor admixture from people whose descendants now occupy far flung parts of the world, but no idea how or when this occurred. Frustrating. I was hoping for more information about my ancient ancestry, not just the last five hundred years as I know that already, but beyond written records. So far nothing.
As far as the health indicators. I did learn some odd things but mostly I knew it already because of my health history or those of close relatives. I consider my SNP Raw Data is unusual compared to others as I am mono ethnic, Southern European and from an island population, and most testees are North American of mixed North and Central Europeans, but it is mostly uneventful. That is the third problem. The lack of a good database of clients who are from all over the world. You can only learn so much from North Americans of mixed European parentage and whose ancestry goes back no more than some hundreds of years.
Ponto,
That’s an interesting perspective. You sound disappointed, which suggests that you expected quite a bit more out of your SNP profile. Looking back, do you think your general optimism led you astray, or do you think 23andMe effectively over-promised and under-delivered?
I don’t know that I got much more than you in objective terms, but I am much more satisfied with the results. Partly that could be from being a (stereotypical) North American of mixed European ancestry.