Genetic genealogy raw data and Alzheimer's research ...

Alzheimer's, cardiovascular, and other chronic diseases; biomarkers, lifestyle, supplements, drugs, and health care.
Post Reply
circular
Senior Contributor
Senior Contributor
Posts: 5565
Joined: Sun Nov 03, 2013 10:43 am

Genetic genealogy raw data and Alzheimer's research ...

Post by circular »

I should first note that this concept is exploratory. I can't pinpoint a specific way in which genetic genealogy raw data from 23andMe, Family Tree DNA (FTDNA) or Ancestry.com (and others) can directly contribute to Alzheimer's research, but maybe we can identify some over time as researcher databases grow and multiply and we learn how to get our family genetics into at least some of those databases.

J11 you honor me voting me thread leader for a genetic genealogy and Alzheimer's research thread. I could try to follow and contribute to such a thread, now here, but I am so so so so so behind in my life, and it's such a big topic, that I'm unlikely to focus on it soon. I recently went on a genetic genealogy unsubscribe binge so I could focus better on all my unfinished projects. A thread like that would require focus on all those things I just unsubscribed from just to get caught up, which I expect will take at least a year thinking positively. But I can at least post some thoughts and we'll see if this thread gets out of the barn. I almost added it to the exome scan thread, but that is very complex and maybe this approach will feel more accessible to people?

I love your enthusiasm for taking a citizen science approach to Alzheimer's genetics. I think though that it's possible that the tools available to genetic genealogists don't hold a candle to what researchers have available to them, and that reasonable inferences concerning genes and health require enormous population level datasets. I'm not sure it will work to start at the family level? But, I think we can make our raw data available to those researchers.

For background, 23andM3 has been primarily in the health DNA business with ancestry on the side, until the FDA made them stop offering the health reports pending regulatory action. Family Tree DNA has always deliberately focused only on ancestry genetics, in part it's thought so as not to get snared in the regulatory issues over health DNA. Ancestry.com started out with genealogy database expertise and later, when they saw what was going on through FTDNA (and 23andMe? - not sure their timing) got into genetic genealogy first with the Y chromosome DNA and later autosomal DNA. Then Ancestry canned their Y DNA testing which really wasn't a good product next to what FTDNA offered.

So now you have the "big three" (among others like the Genographic Project and ...) and each has its advantages and disadvantages for genetic genealogy purposes. All of them let you download your raw data. (Only FTDNA allows someone to import their data from Ancestry or 23andMe into the FTDNA database for comparisons there, free for minimal, fee for all tools.) But autosomal DNA from any of the big three can also be imported into the third party website GedMatch.com to allow cross-company comparison between users to find more genetic matches.

I think the key is for anyone who's had their genetics run at 23andMe, Family Tree DNA, or ancestry, to be sure to upload their raw data to the research outfits if they're satisfied on privacy fronts. This page lists a number of databases out there and notes which ones genetic genealogists can upload their results to http://www.isogg.org/wiki/DNA_databases. Maybe someone will extract those for the different types of DNA (Y chromosome, mtDNA and autosomal DNA) into a handy little list with links.

Ancestry.com recently unrolled a Health DNA arm. See https://health.ancestry.com and the press release http://corporate.ancestry.com/press/pre ... tryhealth/ and this coverage http://www.huffingtonpost.com/2015/04/0 ... 08446.html

Ancestry's effort looks nice in theory, although the ever-present privacy concerns apply. What I think is very problematic in their vision is that a huge percentage of family trees at Ancestry.com have scads of errors. People linked to the wrong family and all kinds of things. Then users process Ancestry's hints like they're eating candy, without critically evaluating them, and misattributions spread like wildfire through the trees. Also people will no doubt enter health information that is not correct, thinking an ancestor had a particular problem when it was something else. If they rely on actual death certificates, those aren't always correct either, and require accurate interpretation of old medical terms. But if Ancestry can leverage their DNA databases around privacy concerns and get them in the hands of Alzheimer's genetics researchers, drug development companies etc., and skip the family tree histories people enter which could be fraught with bad information, then that may be the best to come of it.

J11 keep sleuthing this topic and share here, and of course anyone else. I'll add too if I have new thoughts.
ApoE 3/4 > Thanks in advance for any responses made to my posts.
J11
Contributor
Contributor
Posts: 3351
Joined: Sat May 17, 2014 4:04 pm

Re: Genetic genealogy raw data and Alzheimer's research ...

Post by J11 »

Thank you Cousin Circ.

{After exploring my results on GEDmatch I realize that I am probably related to almost everyone on this planet within 20 generations or so. It sure makes you understand things differently. My family kept family simple. Mother, father, son grandma, granddad, uncle, and aunt. That was about it, everyone else was in the other category. It is going to difficult for me to be grumpy with people on the thread when they can all claim to be my cousin! }

I think this is a great journey to embark upon. Accessing the collective wisdom of our thread and other experts will greatly help us
reaching our shining star.

The idea of seeking out the genetic basis of our loved one's dementia has been developing on the Exome thread for over a year now.
The exome scan returned an enormous and completely unmanageable list of over 60,000 variants. After an arduous search through these
variants, many of these variants appeared to have possible relevance in the causation of dementia. It was not obvious which of these was the actual cause. I sought out a professional medical geneticist for help, though they showed no interest in sorting through all these variants.

Many many people must be stuck in the same situation. The GWAS studies to date have focused on common variants that occur on a global scale. Most of the family context of the variants is lost in this approach. In fact almost 100% of familial AD is still almost entirely unexplored genetic terrain. Our family member with AD while 33 and considered by 23andme to be at low risk for AD has a family history of dementing illness.

My big breakthrough has been my recent discovery of the 23andme family software. I have not paid any notice to these features before.
However, the 23andme website is now listing hundreds and hundreds of distant cousins. Gedmatch has over 1000. We are getting frequent emails from 23andme announcing yet more relatives. At the Gedmatch site when I select one hundred or two hundred relatives and then do a comparison entire chromosomes become covered with shared segments. There is now near saturation coverage of our loved one's genome with shared segments with relatives. We have already reached the time in which tracing through the origins of one's inherited illnesses is possible through a genealogical approach.

It occurred to me that with all these shared segments finding the variant of interest might not be very difficult. This is likely true for almost
everyone else! It is a We the People time! For a long time we have been drifting toward this moment. Now we have arrived! Why have not others noticed this? Almost all the science to this point has been about finding the commonalities of illness at a very broad scale (often a global scale). Investigating the cause of illness in a single family simply might never be a priority in the scientific community.

This is where Citizen Scientists can play a role. The tools are largely already available online to work through the genetics of illnesses at a family scale. Examples of people doing so have already been posted to the web. It would be such a gift to humanity for the broader community to step forward and contribute their genetic knowledge of their illnesses to the collective scientific literature.

This is not a question of if only when. The computer infrastructure and software is mostly in place now. Affordable exome and genome scans are now on the horizon. While, it is not necessary for all of one's relatives to have such scans done to identify variants of interest, they simplify the search. The journey that we are considering on this thread might be a short one. As soon as everything falls into place, the revelation of the origin of genetic illness at the level of the family could occur quickly.

The idea of this thread is to give it a nudge. A little push at just the right point can supercharge change. If we can attract some interest, a whole bunch of interesting things could happen. Scientific researchers do not need that much encouragement to help people out when they see motivated people wanting to accomplish something. I suppose such an endeavor might result in a few publication credits.

I am also sure there would be more than a few politicians that might see a favorable narrative in this idea: Help out ordinary citizens to unravel their genetic illnesses that can cost a huge amount of money to treat after the fact; Associate with the deep emotional resonance of family; Invest probably minimal amounts of money in computer infrastructure to move it forward etc etc. Seems like too good an idea to let go by.

This is a very exciting opportunity!
If we can attract people with a passion to solve their family's illnesses, then we can all learn together.

There are some technical issues that will need some working through, though collectively solving these problems should not be difficult.
For whatever, reason this idea does not appear to have caught on yet. Genealogists are very preoccupied with discovering their tenth generation great grandparent. All those people who have not been very motivated to do likewise might now realize that the tools genealogists have developed can be tweaked to find out why their loved one's are suffering from illnesses today.

Anyone interested in joining us on this adventure?
circular
Senior Contributor
Senior Contributor
Posts: 5565
Joined: Sun Nov 03, 2013 10:43 am

Re: Genetic genealogy raw data and Alzheimer's research ...

Post by circular »

It seems Ancestry.com agrees with you J11 given their new health DNA arm. Another thing to remember is that all this takes place in the context of an uncertain regulatory environment when it comes to direct to consumer genetics. The genetic genealogy endeavors are so far under or outside the regulatory radar, presumably because they're not focused on customers using their results for health purposes. Many would like to keep it that way. If regulators impeded GG companies from allowing customers to download raw data it would put a huge spanner in the GG works as well as health discovery. Just FYI.

There's an ongoing study of generations of a family that's had a lot of AD. I saw it on TV years ago. Googling may pull it up. I don't think they had the early onset form. Not sure.

Are you saying your loved one was beginning AD at age 33???
ApoE 3/4 > Thanks in advance for any responses made to my posts.
J11
Contributor
Contributor
Posts: 3351
Joined: Sat May 17, 2014 4:04 pm

Re: Genetic genealogy raw data and Alzheimer's research ...

Post by J11 »

Sorry that was APOE epsilon 33.

We were surprised. An offspring came back 34. We thought this must mean our loved one with AD must have at least one 4.
Test came back, no 4s. This was very surprising. The other side of the family did not have a prominent history of AD.

It is true about the strategy to keep it genealogical to avoid trouble. However, everything is converging on the family as
the last source of hidden genetic risk. It is the dark matter of genetics. The companies could sell their tests and with good faith
say we are selling you this product so you can find your Great great great .... grand relative, and then people could take
the results and triangulate their results and find out why their loved one's wound up with dementia. Would the government
really step up and regulate the product simply because people were putting the product to an unintended use? It is sort of like
banning computers because they might be used to evade taxes. At this stage it would seem almost impossible to stop the wave
of family genetic research that is now approaching.

And why would they want to anyways? The cost associated with these illnesses is simply astronomical. Informed members of our
forum would readily agree that the Alzheimer's epidemic is rapidly nearing a point in which it will cause the financial collapse of the entire industrialized world. Could government truly be that incompetent to reject the solution to the problem that would otherwise destroy their community? ( Probably best not to answer that question)

Finding the genetics of our family member's AD could have substantial implications. If there truly is a single or possibly a few variants that are driving this problem, then finding them could allow us to prevent anyone in the future to be born with this genotype on a global scale. The dividends for investigating the genetic cause of our family's dementia could be massive.

It is true that some studies have been done on familial AD. However, almost 100% of the genetics of familial AD is unknown. Up to this point, the research has likely focused almost exclusively on large AD families with many affected members. Without the tools that are now emerging on sites such as GEDmatch it would have seemed almost impossible to unravel AD genetics in other families. I had not even given it any thought until about a week ago. We only know of 1 family member with AD: our loved one. We had absolutely no idea how to reach back several generations to other family members who might be located anywhere in the world. The tools that we have recently found appear as thought they could help us and the many others who might be interested to do so.
circular
Senior Contributor
Senior Contributor
Posts: 5565
Joined: Sun Nov 03, 2013 10:43 am

Re: Genetic genealogy raw data and Alzheimer's research ...

Post by circular »

I'm trying to picture this. So first one would purchase autosomal and mtDNA kits for as many family members as possible, to include aunts, uncles, greats, 1,2,3 cousins and so on.

Very accurate family trees would need to be built so triangulation could accurately assign specific shared segments to specific ancestors to be able to correlate segments with family health history.

False matches (identical by state) would need to be abandoned; best to restrict the shared cM threshold very conservatively to give confidence that shared segments are with actual family matches (identical by decent).

Shared segments assigned to specific ancestors will survive only randomly while most of our ancestors' genomes will have been permanently lost through recombination. It seems the chance of family AD genes surviving in shared segments after the random recombination process using partial genomes would be small given the effort.

I'm also still trying to understand family AD genes (bottom up approach) vs those found in GWAS (top down approach). I guess you're saying that the top down approach will miss family-important genes by focusing on the big picture? I'm having trouble understanding why.

I need to really read your exome thread. My gut tells me that learning how to use tools meant to handle much larger portions of our genome than GG companies extract (soon the whole thing), and doing so with many family members, holds more potential?

Not trying to discourage your enthusiasm but it doesn't make sense to me and I don't know if I'm missing something. I think you'd want to work with databases that include the most genetic material and then increase the participant pool size, rather than go with large databases but where they contain limited genetic data for each person; i.e., you don't need as much of one's genome to pursue GG as to find the all important genes for AD, so the GG companies to pay to extract more than you'll need for GG. Of course GG might be revolutionized when we can get our whole genome for cheap. The hard part isn't getting affordable consumer genomes but developing software to enable effective parsing of the data.

Maybe you have a different approach in mind?
ApoE 3/4 > Thanks in advance for any responses made to my posts.
circular
Senior Contributor
Senior Contributor
Posts: 5565
Joined: Sun Nov 03, 2013 10:43 am

Re: Genetic genealogy raw data and Alzheimer's research ...

Post by circular »

J11 a very timely blog post for you:

http://dna-explained.com/2015/09/25/aut ... like-11020
ApoE 3/4 > Thanks in advance for any responses made to my posts.
J11
Contributor
Contributor
Posts: 3351
Joined: Sat May 17, 2014 4:04 pm

Re: Genetic genealogy raw data and Alzheimer's research ...

Post by J11 »

Circ, I am feeling soo positive about this one. Up till now there has not been too much interest in this on our forum because
people with epsilon 4 or 4s have a good idea where there genetic risk lies.

However, half of the people with AD do not have a 4! They are 4 less. This seems like such a great opportunity for the broader interests of the dementia community for these people to find out their risk status. It does make such a difference knowing what your life trajectory is decades before you actually get there. This forum proves this point definitively. I mean what would most of those on the thread right now be doing if I could magically remove their 4s? A bag of chips watching the football game? Our family still has no idea what our risk of dementia is as a result of inheritance from our loved one. It would make a big difference if we knew. It would make a big difference for the other millions of AD families who also have no idea.

This genetics of familial AD idea could double the community of people dedicated to solving dementia. It is the old story when you have
a bet on the table it makes all the difference. Having twice as many people exerting their social-political-economic influence to solve dementia
is not to be dismissed lightly. It could make all the difference.

Remember when I emailed a professional geneticist to help us out there was no response. What are you supposed to do with a VCF file of
60,000 variants? It has taken me a year to figure out that there is no possible way you can find THE variant by simply going through such a file. You need to go the genealogy route. Eliminating large chunks of the genome by simply eliminating regions where relatives did not have dementia makes things much much easier. Every relative you contact will either have dementia (keep that region) or not have dementia (discard that region). Considering the truly overwhelming number of relatives now available through GEDmatch,23adnem and others working this through is now quite possible.

On one of the genealogy sites a poster has done exactly what I am suggesting. This poster was able to determine a rare family variant by using a similar technique as envisioned by me. However, it must have been so much more difficult to do this a few years ago. It is only now that the numbers of people on 23andme and other gene chip services has reached a critical mass. Our loved one now has saturation coverage of their genome. And every month a whole bunch more relatives are popping up. We have drifted into a time in which this idea of citizens finding their own familial genetic risk for illness has arrived. It is just that many have simply not noticed what has happened.

A large part of this is also about cost. I have known for quite some time that if you were to throw thousands of dollars at this then you
could probably work everything out. However, many people simply would not be comfortable spending money in that way. However, with the extensive ancestry that can now be accessed for free, solving unraveling genetic illness could be done for a very reasonable quantity of money. The ancestry sites will gene chip for $100. An exome scan could also be done to narrow done the exact variant within a specific region. This would not be expensive. One has went from a cost in the thousands to a cost probably less than a thousand.

The thing to keep in mind is that for most people with familial genetics there is currently almost no insight as to what the genetic cause is.
This seems very surprising. We do not know what causes dementia in those without a 4? No. Almost 100% of familial AD is of unknown origin. The point to keep in mind is that almost all of the genes found in GWAS studies increase AD risk by 10-20%. Then how could families have dominant inheritance AD? Increasing someone's risk by 10% for each risk allele is probably not going to cause a pattern of AD inheritance. For that to happen you need something big like APOE epsilon 4. You need a big risk increaser. The big problem is that finding such variants at the family level has to this point been difficult. Finding large families of people with multiple affected family members is a challenge in the modern world. In the modern era, you can jump on a plane and arrive half way across the world. After you start your new life how can you find your family back home again? The new genealogy tools give everyone this opportunity.

I have already run this by expert genealogists on threads. They do not deny that what I am suggesting is not possible. In fact, some of them have already done the same thing to solve medical genetics mysteries in their own families. The big problem had I have faced for the last week or so is trying to break through the first generation or two of brick walls. Mostly people talk of these brick walls happening at 4 or 5 generations out. In our family, no one in the near circle of relatives wants to get tested. This is going to be one of the big problems.
Another difficulty is that we are not aware of anyone else in the family (on any branch) that actually has dementia. This seems very unlikely.
Dementia is something most people do not want to talk about. There are some silent branches of the family that will be interesting to investigate further.

A technical point conceded by the genealogists was my contention that when looking for a trait of interest cM do not matter as much as Mb.
For genealogists there is a focus on centimorgans because they are interested in finding relatives within a few generations possibly 4 or 5.
My idea was that I am not that motivated to find relatives as I am to find the reason for our loved one's dementia. For this purpose, it is not as necessary to focus on centimorgans but instead megabases. I want to find matches with the most shared megabases. Centimorgans and megabases are not the same units. The only thing I need to be concerned about is not accepting matches that might stretch back to the cave age because I need the matches to be close enough so that rare variants would not be discarded as the matching conditions became loosened.

So this idea has been given an expert first run through and has been accepted as valid. Yet, there is specific concerns about applying this idea to my particular family.
ApropoE4
Contributor
Contributor
Posts: 396
Joined: Sun Feb 02, 2014 10:43 pm

Re: Genetic genealogy raw data and Alzheimer's research ...

Post by ApropoE4 »

Don't you think 23andMe have already looked at this data? If there were obvious associations they'd be first to publish a paper on it (they're not allowed to offer new health reports to customers, but they're certainly allowed to publish them in scientific journals)
J11
Contributor
Contributor
Posts: 3351
Joined: Sat May 17, 2014 4:04 pm

Re: Genetic genealogy raw data and Alzheimer's research ...

Post by J11 »

I think the big hold back is the phenotype data, especially with dementia.
The APOE report on 23andme is one of the only (the only ?) report(s) which require people to opt in.
It is widely appreciated that a genetic risk report for dementia is a sensitive disclosure for many people.

One of the big pieces of the puzzle has now fallen into place: the genotype data.
This has already made it possible for some to make novel genetic discoveries from their family's pattern of illness.

The phenotype data could lag considerably behind.
It is quite possible that even when genomic technology fully arrives at affordable prices, the complete genetic characterization
of disease might still take quite some time.

One possible way around this would be to report exome sequences for large numbers of healthy people. This would be
be very helpful as it could then help eliminate many of the tens of thousands of variants in an exome sequence that have
no significance.

The big logjam has become phenotype not genotype.
This is where citizens can step up.
The people who truly know phenotypes are families.
User avatar
SusanJ
Senior Contributor
Senior Contributor
Posts: 3058
Joined: Wed Oct 30, 2013 7:33 am
Location: Western Colorado

Re: Genetic genealogy raw data and Alzheimer's research ...

Post by SusanJ »

J, I think you hit the hard part spot on, and that is getting relatives to spit for a good cause. My brothers are not even interested in this stuff.

I wonder if you paid for it, they spit, but the results got returned to you, then they'd be more willing. They don't need to receive the results. Is it a matter of trusting you with the results?
Post Reply