Tuesday, October 4, 2011

openSNP and Personal Genomics

So if you haven't heard, Direct to Customer (DTC) genomics has hit the mainstream. Multiple companies (23andMe, deCODEme, etc.) will now genotype you, providing you with a detailed rundown of all your genetic traits and tendencies...
Well, not exactly.

The promise of personal genomics/personalized medicine is that doctors will someday be able to prescribe medicines and lifestyle recommendations that are customized to how YOUR body works, not to some statistical average of how most studied human bodies work. The rub is that identifying genes that control these individual responses to drugs, foods and experiences is really really hard.

I won't describe it all in detail, but the short story of how personal genomics works is as follows:
  1. Find a big group of people
  2. Identify genetic markers that differ between human genomes (SNPs, currently)
  3. Do math to see what SNP markers are associated with the trait you care about
  4. Check your own genome to see which version of the SNP (and trait) you have
In reality it's not so simple because most traits are controlled by many genes, and many different (but rare) alleles can have the same effect on a trait. This is referred to as the "missing heritability problem." It's not unusual to have an extremely heritable trait (where the value of two parents is very similar to the value of their offspring) yet where (across a population) the presence of specific alleles don't explain much of this association. It's a sticky statistical problem and one that many famous geneticists have declared cannot be fixed no matter how many markers or individuals you add to your study.*

So back to openSNP...
openSNP is an effort to put the power of personal genomics and association mapping (aka linkage disequilibrium, aka GWAS) studies into the hands of the people. The idea is that everyone will upload their SNP marker genotype to this common database (along with whatever part of their own phenotype they want to share) in order to create a shared resource. While published data is already freely available (another site, SNPedia, aggregates these peer-reviewed results), the people behind openSNP protest that databases owned by these personal genomics companies are not available to the public.

I love open source/access efforts and citizen science in general, but they're definitely trying something audacious. I'm not going to underestimate the application of the internet to a difficult problem, but they definitely have it cut out for themselves. I think I'd be hesitant to upload my own genotype due to privacy concerns but it's definitely an interesting project. I haven't considered paying to get my genotype though I did sequence a segment of my mitochondrial genome (which was inherited from my mom's mom's mom's mom back in Abruzzo) while I was in grad school.

Incidentally, I came across openSNP first in the DIYbio google group.
It's worth checking out.



* Association mapping seems to work much better in domesticated plants than in natural populations of humans and other creatures. If you're interested in more detail, these papers are a great start.
** The only reason this works at all is because DNA sequencing has become so cheap that it's already revolutionized how biology is studied. Sequencing costs per basepair is actually falling faster than Moore's Law.

Zhu, C., Gore, M., Buckler, E., & Yu, J. (2008). Status and Prospects of Association Mapping in Plants The Plant Genome Journal, 1 (1) DOI: 10.3835/plantgenome2008.02.0089
Hamblin MT, Buckler ES, & Jannink JL (2011). Population genetics of genomics-based crop improvement methods. Trends in genetics : TIG, 27 (3), 98-106 PMID: 21227531

3 comments:

  1. Thanks for sharing our project here and your ideas! We really are aware that this is a large project that will take lots of time and even more commitment of its users. And i totally agree that, because of privacy issues, this is not something that everybody wants to participate in.

    But there are already so many people making their data available that we want to try :)

    ReplyDelete
  2. I think it's a great idea and I hope you guys are successful. I'm sure there are plenty of academic GWAS databases already, but making one specifically targeted to encourage the public to contribute has tons of potential as a future resource.

    Good luck!

    ReplyDelete
  3. Thanks! There are some GWAS databases but as far as I know them they only list results. So if you are interested in finding all GWAS that checked for SNPs associated with a disease of your choice: You can do this. But as far as I know there are non where you could easily obtain the raw data for case/controls (I'm no real expert on human genetics, so maybe there is already a thing like this of which I don't know about). :)

    ReplyDelete

Share!

LinkWithin

Related Posts with Thumbnails