Monday, January 4, 2010

SNPedia's Top 10 SNPs of the Year

SNPedia now contains nearly 10,000 SNPs and to welcome 2010 we'd like to highlight at least 10. These SNPs have been selected based on an elusive and ultimately subjective combination of medical importance, statistical believability, and overall general interest. This isn't objective science though, so feel free to comment about why your favorite SNPs should have made the list.

1. rs4244285: The antiplatelet drug clopidogrel (Plavix) is the #2 selling drug in the world. This SNP in the CYP2C19 gene can tell you if it's unlikely to be doing you much good; carriers of 1 or especially 2 A alleles don't benefit much in terms of lowered heart attack incidence based on several studies published over the last year. And actually, this fairly common SNP affects how you metabolize plenty of other drugs as well, including anti-depressants and anti-ulcer drugs.

2. rs4149056: Speaking of bestselling (>100 million prescriptions/year) drugs, one of the adverse effects associated with statin use is the possibility of myopathy (muscle problems including weakness, cramping or pain). Carriers of 1 or 2 C alleles for this SNP in the SLCO1B1 gene taking statins are at ~5 or 17 fold higher risk, respectively, for statin-triggered myopathy. Ouch.

3. rs1799853, rs1057910 & rs8050894: How about one more - actually three more - affecting a widely prescribed drug? When assessed together, these SNPs in the CYP2C9 & VKORC1 genes help predict something that's otherwise both tricky and overly empirical to determine: the optimal dose of the anticoagulant drug warfarin (Coumadin ), used primarily to help prevent thrombosis and embolism in patients with high blood pressure. The FDA now recommends - but doesn't require - testing warfarin patients for these SNPs. Here's a long-standing wish, good for 2010 and most likely beyond: Medicare announces it will reimburse for testing these SNPs, given that over 1 million Medicare beneficiaries take warfarin each year.

4. rs10757278: The region of chromosome 9p21 with this SNP (and it's neighbor, rs1333049) has been shown in 2009 by several large studies to indicate at least somewhat increased (~1.3 or 1.7x) risk for coronary artery disease and it's consequences (like heart attacks). But has it added anything to what a cardiologist would already have known based on traditional risk factors like hypertension and family history? Actually, yes - you can classify a patient's risk better when you add the status of this SNP to all the other factors you've already assessed.

5. rs1537415: Sure, lots of SNPs are of interest to MDs, but here's one for DDS's. Over half of us carry at least one allele of this SNP and are therefore at increased risk for periodontitis - so keep flossing!

6. rs3892097: This SNP encodes the CYP2D6*4 allele, the most common nonfunctional variant for this gene. While it can therefore affect the metabolism of about 25% of all known drugs, it's on our list this year due to increasing evidence for poorer outcome among breast cancer patients treated with tamoxifen, the gold standard drug for roughly the last 30 years. Women are typically treated with tamoxifen for 5 or more years, but evidence is mounting that less functional CYP2D6 alleles lead to less endoxifen, which is the active form formed via CYP2D6 metabolism of tamoxifen. Less of the active form means a poorer outcome, so now the question is becoming, can higher doses of tamoxifen overcome this, or should alternative drugs replace it as the gold standard?

7. rs1447295: And now here's one for the guys: this SNP has turned up in studies totaling tens of thousands of patients across multiple ethnic groups as connected to an increased risk for prostate cancer, the second deadliest cancer among men. Although on it's own the increased risk isn't that high (perhaps a doubling of risk even for those carrying two copies), there are now over 20 more SNPs that can be used together to help classify risk. We can hope that more aggressive screening of those deemed to be at higher genetic risk will help decrease prostate cancer deaths.

8. gs138, gs139, gs140: These three genosets (hence the 'gs' prefix) represent the rapid, intermediate, and slow metabolizers amongst you for the detoxifying enzyme NAT2. Debuting in 2009, the algorithms described in SNPedia using seven NAT2 SNPs allow the Promethease software to classify your NAT2 status and thus estimate just how fast you'll clear various toxins (or that hangover?) out of your system. Looking into the 2010 crystal ball: we foresee the use of lots more genosets to predict phenotypes (and ancestry) based on combinations of SNPs.

9. rs17646946 & rs11803731: Two SNPs, both on chromosome 1, yet 27 million bases apart; one is part of the trichohyalin gene, the other abuts a trichohyalin-like gene. One found by a genome-wide association scan by academic researchers, the other by a company harnessing the power of the internet , DNA chips, and customer interest in correlating their DNA with common traits - in this case, hair curliness. We applaud the use of people-power to help prove (and disprove) DNA associations, and in 2009 we began having SNPedia/Promethease users self-report their own associations for a variety of SNPS, regardless of which company they got their genome data from.

10. rs2395029: This SNP just keeps getting more associations every year! As the SNP predicting the presence of an HLA-B*5701 allele, it's been previously associated with a variety of conditions (like psoriasis, or as the FDA agreed last year, abacavir hypersensitivity). This year, it demands attention for the remarkable increase in risk (45 to 80 fold) for liver damage among patients taking the antibiotic flucloxacillin (aka floxacillin). While this SNP is quite rare, it's a good example of a more "deterministic" type of SNP that we can carry.

On that last note, the intersection of less expensive full genomic sequence with greater coverage of rare variations bodes well for all. Think well of all the researchers who make this possible - we thank you for your hard work, we encourage you to choose open access publications, and we look forward to all of your contributions in the coming year!

2 comments:

dan said...

Very impressive list!

It's a really great idea to try to make this kind of information more accessible (and open) to the public.

Have you thought of trying to get the X-Men interested in sponsoring SNPedia (or there Marvel spokesperson at least)?

cariaso said...

At present dbSNP has poor coverage of super power causing mutations, since so many of them are actually epigenomic or based on radioactive chimeric genomes. However I've heard that OMISH (online mendelian inheritance in super heros) has been funded for next year. If so, the next challenge is finding sufficient numbers of volunteers for the superhero/supervillan GWAS.