snpedia.com is 2 years old today. olympic dreams to all. especially the burmese.

0.1.24 public genomes

Promethease 0.1.24 has been released.

A growing collection of public genomes allows you to see how magnitudes have improved the system.

Mendelian Inheritance

Promethease 0.1.8 now available with an important new feature called a genoset. Genosets allow us to talk about specific combinations of SNPs.

As an example: 77% of Europeans who lack the rs4988235(T) and rs182549(A) haplotype will be lactose intolerant.

Previous releases could show you had the high risk form of each, but it couldn't highlight that you actually had both. Genosets resolve this. Each genoset gets a unique gs#, similar to an rs#. It also has a criteria which must be met for you to see the consequence. It works like the connection between a genotype and a phenotype.

Here is the genoset which predicts lactose intolerance and it's criteria
http://www.snpedia.com/index.php?title=gs100
http://www.snpedia.com/index.php?title=gs100/criteria

A criteria is a little computer program which evaluates to True or False for your specific genotypes. If True, your report will show the text from the main genoset page. Relax, only hardcore SNPedians will ever see another criteria.

Here's what the promethease reports look like for 23andMe's demo user Lilly Mendel and Greg Mendel.

Lilly Mendel should have no problems with lactose. Greg Mendel is probably lactose intolerant and may also be at increased risk of ALS. But look just below that and you will see that both Mendels carry a rare rs4253208(G;G).

Here from 23andMe's original file is the raw data

rsid	chromosome	position	genotype
rs4253208	10	50348728	GG

And NCBI/HapMap shows that the the (G;G) genotype is possible but rare. In fact it even suggests the Mendels may be closer to African than European ancestry?!

What's actually happening is strand confusion. SNPedia is oriented to NCBI dbSNP. 23andMe probably oriented to the opposite strand. However, there is nothing in their data file which indicates this. Without it, promethease has to recognize the (G;G) as possible.

deCODEme's file format does pass strand information, but perhaps the time has come to consider what a long-term personal genomics file format should look like. XML is well suited to this. Strand information is clearly important. I'd discourage 23andMe's fondness for pushing both allele's together to avoid confusion when dealing with multinucleotide snps such as rs34815109. Ideally it would also be able to handle sequence information since microarrays can't really see haplotypes.

Lastly, several people have reported seeing the message 'Not a win32 application' when trying to run promethease. It seems all were caused by failed downloads. If you see that message please ensure you have the full 7M file from
http://www.snpedia.com/files/promethease/promethease.exe

Promethease deCodes 23andMe

Promethease annotates your genome using SNPedia.


It is possible thanks to the Personal Genome Explorer which allows 23andMe customers to export all of their 550,000 genotypes. Promethease reads this export format, and uses it to generate sample reports.


Your genotypes will probably be a bit more boring, since these random files get some pretty weird diseases. The report format isn't yet as intuitive as it should be, but you'll manage.


Your changes to SNPedia will appear in future reports. To generate a report simply drop an export file onto promethease.exe . From the DOS command line you can compare to other hapmap populations, instead of the default CEU. Runtime is about 2 hours. It doesn't eat much of your net or cpu, it just moves slowly to be courteous to others.

PGE does not yet provide information to orient snps relative to dbsnp. As a result many of the genotypes are flipped. This causes Promethease to ignore population frequencies. These SNPs will show with a 'None' along their left side.

http://www.snpedia.com/files/promethease/promethease.exe

GBrowse SNPedia

It is now easy to overlay SNPedia on the Gbrowse databases. This means you can generate images like

(click for full)


which shows a cluster of snps around TCF7L2. The upper track identifies SNPedia snps. The the lower track identifies all snps present on the Illumina 550 microarray. All are hyperlinked to their SNPedia pages. If you click on the image above, you'll see the full image with the LD plot at the bottom. This allows you to determine how much two neighboring snps can say about each other.

To generate your own visit the hapmap of Chr10:114724078..114764077

At the very bottom of the page add these two 'remote annotation urls'
http://www.snpedia.com/files/gbrowse/snpedia
http://www.snpedia.com/files/gbrowse/microarrays
The snpedia file is a modest 300kbytes, but the single illumina chip is 58Mbytes!

Every snpedia rs# page now has a direct link to the hapmap site. Loading will take a while. Scenic spots along the genome include:

rs7901695
*click on hapmap
*set Scroll/Zoom=40kbp (upper right)

rs9939609
*click on hapmap
*set Scroll/Zoom=100kb (upper right)




I'm aware of a problem with rs3135506, but unable to explain it. Other times, snps listed in the snpedia files files aren't rendering onto the images. Other bugs probably remain. Your bug reports are always helpful.


Annotations will be updated approximately weekly, along with the fasta.


(click for full)

Fasta format available

The fasta formatted version of all the SNPedia snps is now available at
http://www.snpedia.com/files/snpedia.fasta.gz
The above link will be updated regularly.

SNPedia was mentioned in the current issues of Nature [1] and Science [2]

Venter’s Genome: No Genome Is An Island

With the release of the research paper describing the personal genome of J. Craig Venter, ostensibly the first individual genome to ever be made public, we’ve got an incredible opportunity to muse about what we know so far about human genomes. And the answer is: not much today, but wait until tomorrow.

I don’t know about you, but finding out that Venter has the ear-wax characteristics of most Caucasians and has European ancestors isn’t that surprising to me. His risk for heart disease and Alzheimer’s was already known and could have been found out a lot quicker and cheaper using microarrays. And although we believe that in these early days of personal genomics, you haven’t really released all your sequence unless you’ve released the primary sequence reads with their quality score files, the release by Venter of his assembled sequences (even without the primary data) does at least allow some geno-dumpster diving. We’ll probably post a few nuggets of information about Venter that aren’t mentioned in the PLoS article in future blogs, but suffice it to say for the moment, Venter is mostly pretty normal for a white guy.

So where’s the beef? Think about any nascent network. Who did the very first people with telephones get to call? The first people establishing a network get – initially – very little benefit, aside from the well-deserved credits. The value to scientific research? Priceless. The value to the individual? Hardly.

Venter reports a lot of novel genetic variants, including copy number variations, deletions, insertions, duplications, you name it … and of course, since they are by definition novel, we don’t know what they mean other than at least they aren’t fatal or associated with obvious overt disease(s), given that Venter is alive and reasonably healthy.

For all the hoopla over the Human Genome Project, the reality is that we still know very little about the complex interplay between millions of genetic variations and environmental factors and lifestyle choices. It will take many, many individual genomes, and even more daunting, detailed comprehensive medical records, before we are able to make many of the correlations that will ultimately matter to us on a daily basis. Venter estimates a minimum of 10,000 individual genomes, and maybe, a million or more.

Let me put it another way. What’s it worth to you to get your own genome sequenced, compared to just using microarrays to determine your genotypes? Here’s my first rule of thumb: it’s worth one penny times the number of already released complete personal genomes that also have medical histories. So with Venter being #1 (OK, he actually hasn’t released any medical info really, but let’s cut him some slack), in just another 9,999 genomes I’d say your personal genome sequence will be worth about $100 more to you than whatever you could learn from genotyping yourself with microarrays (and using SNPedia, naturally).

One corollary to this is that it will be essential to get as many genomes sequenced as possible, and whether it’s from government support or celebrity genome sequencing shouldn’t matter. What matters is that the medical histories of the individuals are also made available. As is the case with genome association studies, both case-control (groups of patients with a disease; why not one of the patientslikeme.com communities?) and ‘genome cohort’ personal genome sequencing studies should be funded by whoever’s got the bucks. Comparative personal genomics, here we come!

I recall a joke that probably plenty of folks have told; I heard it from Francis Collins, the head of NIH’s Genome Project.

A previously-married woman heads to bed for the first time with her new beau, and to his surprise, she admits to being a virgin. When he wonders why, she says, “Well, I was married to a genome biologist, and every night, he just sat in bed and talked about how great our sex life would be someday.”

Someday personal genome sequences will be great for each of us to have, and it’s fantastic to have the first personal genomes coming out this year. As of today, it’s 1 (personal genome) down, at least 9,999 to go …