23andme comparison for people with ME?

[snowathlete]

A number of people with ME have had 23andme tests including me. I wonder if there is some way in which we could build a database of results and see what the vast majority of us have in common?

I mean if it is genetic in some way then we should have common genes...right? I'm guessing that we would be looking for what we share in common then compare against healthy controls and see what is left at the end as unique about our patient group.

Anyone know how we might achieve this?

 

[searcher]

I don't remember who owns this spreadsheet, but the easiest way would be to build on the spreadsheet at https://docs.google.com/spreadsheet/ccc?key=0Ar76dNWyEQLIdExKZUhfXzFUWHpkWnZObk1TM1dxX0E&hl=en#gid=0 Does anyone know who owns that google spreadsheet? It looks like have found some significant differences compared to controls with 49 CFS patients, and since so many of us have done 23andme now I bet we could find additional statistically signifcant SNPs.

If we wanted to look across all of the SNPs, a more complex way (but not that complex) would be for someone to gather the 23andme exports and do some analyses. It would be easy to anonymize since you can remove any personal information from the export.

 

[snowathlete]

The exports are just text files right?

Anyone have any programing skills?

 

[searcher]

The exports are just text files and are easy to read. I am a programmer although I have only taken one class in bioinformatics. I talked to RIch a couple of years ago about doing a project collating and interpreting 23andme data, but I wasn't doing well enough at the time to take it on. I think kday has been doing some work on parsing the 23andme files so he may be the logical person to lead such a project if he is up for it.

 

[kday]

Hey, I've been a combination of busy (taking on multiple projects at once) and sick, so I admittedly haven't done much with the site. I didn't think I would need help on the programming side, but maybe I do if we want to get a database like this working.

It wouldn't be too trivial to put everything in a database and find out what SNPs we have most in common. The hard part would be organizing the SNPs by how rare they are, implementing a database of which are the "good" and "bad" mutations, and parsing the data into something meaningful besides a bunch of rsid's. Unless somebody already has already organized a database with this information.

Most likely, I would just have a survey before you upload your file (to sort files by Autism, CFS, etc), and just provide anonymous raw data for anyone to download analyze. And they can use their own programs/tools.

 

[kday]

Unless someone else knows how to integrate all the data I mentioned

 

[kday]

But if there are people out there (researchers or hobby researchers reading this?), I can collect data and provide downloads for any data people submit. But if nobody is interested in this data, I'm not going to waste my time providing this info.

I currently don't collect any data, and I would give the user the option whether or not they want their data collected.

 

[caledonia]

Curetogether.com has recently teamed up with 23andme.com (about August 2012), so at some point they may release data on SNPs frequency and various conditions. CFS is one of the health conditoins on the site. I would guess they're probably waiting until enough people upload their SNPs to be able to make a report.

 

[kday]

For programmers: snp135.txt.gz looks like the file you want for frequency of SNPs, etc. It's a 1.3 gb file when downloaded and about 9 GB text file when unarchived. It can imported into a sql database.

I'm not going to do it, but you could strip out all the SNPs that aren't in 23andMe and you'll have a nice condensed database of everything you need to find the most common rare mutations, etc.

http://hgdownload.cse.ucsc.edu/goldenpath/hg19/database/

 

[Harry35]

I see 2 difficulties with this. First, the latest (V3) chip used by 23andMe has almost a million SNPs. That's a lot. Are we going to compare all million SNPs? How big a computer would that take? The google spreadsheet above only looks at the 31 Yasko SNPs. You can't find new SNPs that correlate with methylation from that list, we already know that all of them have some influence.

Second, how will the people in a given category (ME/CFS, for example) be determined? Do we trust self-reporting? I've got some methylation problems, but I've never been diagnosed as ME/CFS or one of the related syndromes. My problem is migraines, which correlate with methylation mutations to some extent.

Actually, 23andMe has the archives of all the tests they have run, but they seem to be hesitant to do much with their data if it hasn't been thoroughly vetted. That's why their analysis of each person's probabilities is so skimpy. They don't even look for MTHFR issues, for example. I'm compound heterozygous for MTHFR, but they didn't flag that to me.

 

[snowathlete]

I'm not an expert but i'd have thought that the data size is still managable, though it would take a while for a computer to do its stuff. It's not a task for Excel.

I'd suggest limiting categorizing by various levels, so that you can filter and analyse by clinical diagnosis against various different criteia (Internaitonal, Fukuda, Canadian, etc) self-diagnosis, non-diagnosis. Many people might sit in several different camps, like me, I meet several different criteria for ME and CFS.

This might actually be a good place to start; see what paterns present themselves between these different groups. If any.

 

[Anteah]

Hi, I already spoke with Kday about possibly using the data if and when he begins collecting it, but I was wondering whether I should make a separate post and just ask the people on this forum who have their 23andme results whether they would be interested in supplying me with their data (minus the identification info) for a comparison of their SNPs info with a healthy sample. I myself have a CFS but a light form of it, yet also could qualify for ASD diagnosis.

How I got interested in the topic is I noticed that many people in my family share similar traits and that is what first gave me a strong clue that the root of our predispositions and sensitivities is likely genetic, but until now I had no opportunity to study it. My background is in IT (over 10 years of experience) and counseling (I have an MA in Psychology), so as you can see I am not a health professional or a researcher by trade, but I am very familiar with reading various research studies and have taken some graduate level classes in research, so I am hoping that slowly but surely I can hack through the data in hopes of finding some patterns that we CFS/ME folks share and that separate us from the sample who do not present with our symptoms. I have basic knowledge of programming and databases but am able to summon up all the help I heed to get this project off the ground.

Obviously it is a very grassroots kind of undertaking, but I do have a strong interest in making this as reliable and scientifically valid as possible. Where I will be out of my depth I am able to engage others with more experience in the subject such as my friends who have programming or research experience. I am also hoping that if I do end up being provided the sufficient data that I can implement the suggestions of the forum members to insure that the results are those that are able to answer the most pertinent questions the majority here is interested in. Another words it is all of ours effort, I am just able and very willing to give it all the effort it needs in terms of initiation, back office work and funding if and when needed.

I guess my question is what are your guys opinions, is there a need for such study? If say I post a separate thread asking folks to provide me with their data will that be appropriate? Should I clear it first with the moderators? I am very excited to be able to study something that I was just able to wonder about for many years, but I also do not want to do something that will go against the grid of this community. So any and all suggestions are welcome. Please let me know what you think about this. Thanks!

 

[Aileen]

It's great to see someone so eager to do such a big undertaking!! I would certainly be happy to provide you with my data as long as there is no identifying info with it.

 

[Anteah]

Thanks so much, Aileen! Yes, definitely no identifying info is needed. I am thinking to create a small website front where I'd explain what this study is about, what are the expected outcomes, how data is collected and will be able to post updates, give my contact info and have an upload link for those generous souls willing to donate their anonymous data to the cause.

This way it will be a little bit more efficient. If you feel more comfortable waiting until i do that just let me know, or if you are ok with just sending it in email thats even better because this way i can slowly start working on the foundation of the application. So far I have two files mine (CFS) and my husband's ('healthy'), so a third entry is a good start . If email is ok i can send you a message with with it, since I am not sure if I am allowed to post it here.

Also anyone else who is willing to help out with data, please let me know, and I greatly appreciate it.

 

[Aileen]

You're going to have to wait a bit for my results. I'm still waiting for them myself! Hopefully I'll get them next week (fingers crossed).

 

[Anteah]

It's no problem, just let me know. I should hopefully have an upload site by then anyway.

 

[snowathlete]

Anteah, thanks for volunteering. I'm very willing to provide my results. Its probably worth another post asking people to take part. i dont imagine that woulld break any rules. id suggest getting a site up first with some blurb.

 

[Anteah]

Thanks for suggestions, snowathlete, and for willingness to share your data! I will make sure to put a site together shortly and post a separate thread. Hopefully people will respond favorably, because to make it as scientifically significant as possible we will need a good size sample.

 

[martiz]

I would be very interested in helping with this project. I am currently over-extended in my volunteer work but could try to finish up some things.

I just got my 23andme data and have mapped it using the common spreadsheet that is out there. I would like to do the same for the genes mapped by Kerr's 2005 and 2008 studies.

But the project you are suggesting is a bigger step forward.

Marti Zavala

 

[SickOfSickness]

I think what we need is to petition 23andme to have more focus on ME/CFS and ME/CFS symptoms. And more ME/CFS patients who opted in to let them use the data and answer 23andme's relevant surveys. 23andme has the ability to analyze so many patients at once, but they are working on so many illnesses and symptoms, if we can get them aware that our patient group is important, I feel that's our best hope for genetic comparisons. They have the technology and volume.