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

[J11]

Hey circ, yeah I'm feeling mighty toasty nice about this whole deal now.

My contacts seem to have taken a 2 month wine tasting tour of the South of France. {What is wrong with people today? No possible way in the world that I would recline on a beach or in the countryside of Provence without restraints and heavy duty sedatives. What are you supposed to do?} Looks like I'll be able to hit them up for the first class genomics tour, "Yes step right this way monsieur, why yes we can set you up with an exome scan, oh did you say you want an amyloid scan, but of course".

With an exome scan it would not even matter anymore about the cM size of the segments! We could be down to single common variants! This would be amazing! There would be no worry that variants would have somehow gone missing. I would have to write a little program for this, though that should be fairly simple. Only tricky part would be those variants that did not have an rs number and the base position changed through different versions of the reference alignment.

Using this approach the only escape route to avoid detection would be if the variants were to go extraexomic. Well, in such a circumstance we would have at least some idea of the general region with the gene chip results. Nowadays the way things are these cheeky little variant devils might find some way of getting down to the general region of a certain wine growing nation of Europe!

[circular]

You're way ahead of me!

[J11]

ICAD2016 in Toronto has an interesting report out on AD genetics.
As we have seen lately, there appears for some unexplained reason to have been very little research into variants that
protect against AD. It is very difficult to understand why this is true. For some reason the whole effort in the genetics of
AD has been devoted to finding bad news. There is as we have seen also quite a bit of good news.

One piece of good news that is now being reported is for Rab10 (page 4 of the pdf).

"Significant association in the protective direction was found for variants rs142787485 (p=0.018, OR=0.58) and rs7653 (p=0.0049, OR=0.35). RAB10 overexpression resulted in a significant increase in the Aβ42/Aβ40 ratio (p=0.017) and RAB10 silencing
resulted in a significant decrease in the Aβ42/Aβ40 ratio (p=0.048). "

http://www.multivu.com/players/English/ ... 562930.pdf

Checked our loved one's exome file. And yes we have the xxx7485 variant!

I went to dbsnp and looked up the location of 7485. 7485 is located on chromosome 2 at 26.4 MB
(as per the exome file and 26.1 MB as per dbsnp) which might help explain a few things.

Whenever I contact relatives who are on the chromosome stretch between 28 and 45 MB I can be certain that
they would not have had a history of dementia. No where else in the genome has so consistently and emphatically
reported a negative history of dementia. I have thought for quite some time that there must be a protective variant
on this segment.

The common shared segment is just outside of the RAB10 variant so it might not be RAB10 (would want to double check
that there have not been numbering changes etc.). There is always the possibility that another another protective variant
lurks along the 28-45 MB stretch of chromosome 2.

All this makes me wonder whether there might be a large genetic risk somewhere in our loved one's genome. 7485 appears to
offer substantial protection against AD. There could be a big one out there that counteracts this protection.

Checked the Broad.
There are no good proxies for 7485.
It would be so great if gene chips could be designed so that all SNPs could be captured by a least one good proxy.
Admittedly to do so might require a gene chip with 100 million SNPs on it.

[J11]

Wanted to make a note of 2 variants that Mutation Taster is calling disease causing in the UBQLN2 gene on chromosome X.
UBQNL2 is apparently involved in many neurodegenerative illnesses.

If males were to inherit this mutation from their mothers, then would have only the "bad" version of the gene. I am not sure whether X chromosome inheritance has been much discussed in AD research, though this could be quite scary.

Oh, and these variants are high quality though apparently are not reported to dbsnp.

[J11]

Hmm, I'm just trying to get a new computer up and running with Genome Browser.
Anyone know how to upload a fasta file into Genome Browser?

I have started to get things going with the VCF and some annotation tracks.
Things just get better and better, Browser allows the Exac fileset to be added in.

It appears that the 2 UBQLN2 variants are on the ends with a whole bunch of ALS frontotemporal
dementia variants in the middle.

[circular]

Wish I had time to do the kinds of things you're doing and help. This group isn't very active, but you may find that someone on it can help with the software questions.

[J11]

Got Genome Brower up and running!
Huge download needed of about 8.5 GB.
Had to do without our monthly fix of cartoons.

Also just got Illumina's Variant Studio going today.
I was getting an error with this program for about two years and today I finally got it to function.

Typed in Alzheimer and the SNP that went to the top of the list was in the MAPT gene.
It's hard to know whether this variant is what I'm looking for, though it is scary enough.

There are several SNPs close by that cause Pick's and frontotemporal dementia.
One problem is that the sequencing quality is a little weak.
This might have to do with the fact that there is a massive stretch of consecutive A's.

It would be almost too much to take, if they have now cured Alzheimer's with aducanumab only for
us to find out that it wasn't exactly Alzheimer's at all.

Mutation taster is calling the variant disease causing.

I'll see if other DNA relatives on this stretch also have dementia problems.

[J11]

I want to get everyone up to date with the search for the variant.
There has been a development and I would love to hear comments from the forum.

About a year ago I contacted a relative on 23andme and asked about their family Alzheimer status.
This was just an out of the blue random match.
It turned out that this match who is in their mid 60s already was experiencing cognitive concerns and both of their parents developed AD. This match at the time did not seem that out of the ordinary from many other matches that I had randomly contacted. About half of them have some dementia in the family, so I did not think this match to be all that atypical. {However, when I have given this more thought lately, I realize that yeah
both parents with AD and a child developing it in the mid 60s is of note even in my matches. This match
would be within the top 5-10% in terms of AD risk burden. I have tried to reestablish contact with this relative though I have not heard back. I am afraid that the impairment might have significantly increased over the last year.}

As I noted earlier on this thread I recently got Illumina's Variant Studio up and running (took me two years!). When I uploaded the VCF file from the exome sequencing I received about a dozen genes. One of these genes was APPB2. I checked the file and I did not find anything of interest in the exome file. Everything looked common, nothing seemed remarkable.

A few days after I ran the software, a DNA relative contacted me and mentioned that they had dementia in their family. This was present in a few relatives including a grandmother. At this point I became intrigued and I put together all of the above into a hunch that maybe there was something to this APPB2 Chromosome 4 40-50 Mb segment. The two matches triangulated with my loved one.

I then went to GEDmatch and started contacting those who were also on this segment. I started receiving quite a few positive responses. The father of one of the matches was exhibiting signs of at least mild AD at age 60. Another match was also in their early 60s, had a strong family history of AD and was self aware of signs of perhaps MCI. Other of these relatives on the segment had a strong family history of Alzheimer's though had not exhibited signs of impairment past the age of typical onset. Others had perhaps some impairment at an age somewhat past what would be expected, though they did not want to label it as Alzheimer's. They thought of it as normal cognitive aging. A few had no known Alzheimer family history and were not exhibiting signs even into their mid 60s, so they felt they were all clear.

I also contacted relatives on this segment who shared the same strand as those with dementia, though they
had not inherited the small segment at the start which included the APPB2 gene. None of the three that responded had any family history of dementia, nor did a relative who shared the same segment yet was on the wrong "strand".

All together there are about 20 people that I have contacted
and most of these fit the theory: they (or more broadly their family) should have developed AD and did, they should not have developed and did not (4 are in this category), or they should have developed dementia, though they did not, yet this can sometimes be explained (e.g. by APOE status). I have not settled it out with everyone, though if I can contact a few more matches about their APOE status, then I might be able to resolve all contradictions.

There seems to be a substantial amount of earlish onset AD in these matches.
I have narrowed it down to a very small segment of DNA of about 1 quarter million
base pairs around 41 Mb on Chromosome 4 in the APBB2 gene which seems to be causing
the trouble. Obviously the fact that this makes no sense from what the exome sequence is reporting
is not encouraging.

Clearly when making a prediction it is much better to be as vague as possible.
"There will be a mighty wind from 25.15 degrees East of North at 1:15:00.0 on ..."
would not be considered advisable by the fortune tellers.
"There could be a wind or perhaps a light breeze from the North, South, East or West
soonish..." is much better and I could predict with confidence that such a prediction
would be neither refutable nor not refutable.
When you move the yardsticks down to a quarter million base pairs, then it is either right
or wrong.

There were not that many outliers to the idea that this segment had a dementia risk enhancer.
One of the more glaring outliers really did not make sense. The person lived into their mid 80s
was on the risk segment and strand and had no impairment. The person had no siblings and no
family history. I contacted the family and asked about their APOE status. Epsilon 2,3. That could
explain things.

Another piece of wiggle room that I found was that these relatives also shared smaller cM segments.
One of them had over 10 smallish shared segments down to 3 cM with our loved one. Typically I would
discount such segments as noise. Usually you are looking for a 7 cM segment for it to be regarded as a valid
match. However, all of these matches had a large shared segment of Chromosome 4 typically of 10 cM.
It should not be unexpected that true matches would also have smaller splicing regions in common. In fact some of the matches triangulated to these small matches.

One of the matches matched on the segment which included the UBQNL2 segment where our loved one
has a very rare variant as reported by the exome file.

What I feel particularly good about with this possible risk match on this segment is that it can be further
investigated. The most recent common ancestor lived about 5-7 generations ago and a few of the matches
appear to have identified this ancestor and have a way to reach the MRCA. If there is truly an AD risk
increaser here it should be findable. If valid, it would be expected that there should be a substantial number
of lines that could be found with affecteds.

[Sandy57]

J11 you are awesome. That is a lot of work you have put into this. Good luck you deserve it, I don't think anybody could work harder to help their loved ones.
Frank

[circular]

APPB2 ... not to be confused with amyloid precursor protein gene which is connected with early onset AD. APPB2, on brief glance appears linked to Bacillus paralicheniformis ATCC 9945a DNA in ours, and connected with purine metabolism? Do I have this totally mixed up or does this relate to a relatively recent article talking about how much viral/microorganism type DNA we have in ours? Sorry not much time to dig, but fascinated by your process!