Can we identify the SNP associated with "What color is this dress" by today?

Question

[Giovanni M Dall'Olio]

You have probably heard about the story of this dress, that some people see blue, and other white.

My question is: can we identify the SNP associated to it, by the end of today?

My idea is:

• Go to 23andMe, and ask people to answer the question and to provide their genotype data. I did it already, but I am not sure if everybody can access it (https://www.23andme.com/you/community/thread/35598/)
• where else can we get genotype data and answers to the question? Are you aware of datasets like 23andMe, with a responding community and genotype data available?

---

UPDATE

It seems that 23andMe added a survey on the color of the dress, so they will probably publish a report in the close future. Unfortunately they will never release the data, nor allow anyone to work on it.

ongenetics initiative

There is a nice alternative initiative from the ongenetics blog, where they collected family trio data. This will allow to determine whether the trait has can be associated to a mendelian trait:

• http://ongenetics.blogspot.de/2015/03/no-simple-genetics-for-thedress-but.html

openSNP

There are currently 4 genomes in the openSNP initiative. I'll soon add mine to make it 5 ;-)

[Konrad]

I’m assuming this is no more genetic than the direction of the spinning dancer. There may be a very slight predisposition, but since the colour of the dress changes for people from one time to the next (e.g. for me) it would be hard to measure.

[Jonathan Dursi]

It would also probably be interesting to find out what other loss of function is suffered by the poor souls who mistakenly see this dress as blue and black.

[Emily]

Just commenting so I get notifications on this. Let us know when you find an answer – those of us without 23andMe can't see the thread on there.

[Giovanni M Dall'Olio]

Thank you for commenting. At the moment the 23andme post doesn't contain any answer from the staff, just a lot of people saying which color they see ;-)

[Peter]

See

Logged in to @23andMe for the first time in forever to answer a survey about #thedress #TeamWhiteGold pic.twitter.com/qtcNBa6ptD

— Andrew Su (@andrewsu) March 15, 2015

- seems they have added this now

[Giovanni M Dall'Olio]

Yes, they added the survey already. But we failed to identify the associated SNP in one day! :-)

23andMe will never release the data, nor allow anyone to work on it. However, there is a nice initiative from the ongenetics blog, where they collected family trio data: http://ongenetics.blogspot.de/2015/03/no-simple-genetics-for-thedress-but.html

Answer 1

[Mary]

That is so nerdy. But I will admit I was thinking about my genotype. My brothers are red-green color blind, and I expect I'm a carrier. I wondered if that status affected me.

Anyway, my data is in 23andme and I'd answer. But you won't recognize my name there. I submitted with a fake name.

[Giovanni M Dall'Olio]

Thank you for answering! Unfortunately only the 23andMe staff could really use of their data. I was trying to convince them to do a quick survey on their clients and provide a raw estimate of which SNPs could be interesting. The 23andMe data would be great because they have access to other type of information on each individual, e.g. eye color, environment, etc.. Unfortunately I don't know anybody at 23andMe :-)

[Mary]

They responded very quickly to me on twitter once. And I just saw them post something on G+. I will try to get their attention on G+, you try twitter.

[Giovanni M Dall'Olio]

Done, I twit!

[Mary]

FYI:

I wonder how #thedress illusion might be inherited. Do you have phenotype data for a family? http://t.co/tDCBXReZV6 http://t.co/6JpCwIyjwX

— Steve Mount (@ongenetics) March 1, 2015

[Peter]

Nice idea - a self contained family history questionnaire could provide evidence of heritability of this phenotype (without needing the SNP data from 23andme).

[Giovanni M Dall'Olio]

Nice idea, thank you for posting it! Unfortunately with this approach you can only infer whether the trait is potentially genetic or not, but to identify the loci associated we would need genetic data (which would be tricky to collect, for ethical issues). In any case, it is worth a try.

[Peter]

Yes, unless 23andme get involved there is no point collecting responses on that thread because you'll not be able to analyse it without the SNP data. If 23andme could set this up as one of their research questionnaires then it could work...