Exome Scan has arrived! Help!!!

[J11]

Using the annotated file from the app from basespace makes moving through the exome quite easy. Simply searching for "Alzheimer" is useful. When this was done the Hemochromatosis, Prion Protein / Jacob Creutzfeldt, Bleomycin, Sortilin1, ABPP4, Plasminogen activator, NOS3, and A2M AD variants popped up. Some of these variants sound scary (especially the Prion Protein/ Jacob Creutzfeldt one), though only the HFE is rare and none of them is likely the variant of interest. So, in terms of the above board read out, there was nothing much to report.


The underground search found much more interesting results. When I used the search word "amyloid", the genes/diseases
Amyloid beta A4 precursor protein-binding, family A, member 3
Beta-site amyloid beta A4 precursor protein-cleaving enzyme 2 (Down syndrome region aspartic protease)
Amyloidosis, primary localized cutaneous, 1, 105250 (3)
Amyloid beta A4 precursor protein-binding, family B, member-1
Beta-site amyloid beta A4 precursor protein-cleaving enzyme (secretase, beta; memapsin 2)

had variants.

I also noticed that there were quite a few BACE2 (BACE2 beta-site APP-cleaving enzyme 2) mutations and there was a full page of spinocerebellar ataxia mutations.

The genetics community is struggling with the rare variants found in some of the above genes. All the genes with amyloid or BACE2 gave me a very bad feeling. If there were a mutation in a gene that cuts amyloid that obviously could be a concern. There were many other scary mutations with the words microtubule, Sirtuin etc. . You just need to have some recognition of these words to know that if they were malfunctioning that this could have significant implications for AD disease risk. Many other genes with bland unobtrusive names might also be involved in AD pathogenesis: they just did not seem scary.

I wanted the members of the forum to realize that the really interesting variants are likely still beyond grasp of current technology.
The variants that occur at a frequency of .1% or less were not on the commercial gene chips and are only now accessible to the research community through full genome scans. The recent deCODE study with 2,600 full AD genomes is the next wave of genetics research. These studies will finally unlock the genome and make analysis of a genome routine. As it is now, an exome
will contain thousands of "variants of unknown significance".

If you wait a year or two, this likely will all be sorted out. At that time a $50 Alzchip might be able to give you a precise risk assessment without the stress.

[J11]

Thank you everyone for the kind words.

I just want to stay away from any blame if you were to drop some large money on a scan and then be disappointed.
It looks as though the analysis buildout is still ongoing. Everyone can see that the large new industry of genomics is now emerging.

It might take us multiple rounds of sequencing for our loved one (and possibly other family members) to figure out where the AD risk variants are located. We would do this right now, if it were not for the considerable expense that this would entail.

[J11]

Here is an update on the exome scan analysis.

I have finally went commercial with the exome scan. I am not cheap just careful with money.
I have tried out Omicia's Opal program and it is one of the better exome analysis programs I have tried. Most of the other
programs I have tried do not consider the question of homozygous/heterozygous/dominance/recessive. This leads to considering a large number of variants that are likely not important.

In addition to the Opal program, I have also found the Exact exome database. This appears to be a very helpful resource. The Exome Variant Server (EVS) was the exome dataset that I consulted when considering exomic variants. The EVS consists of about 4000 exomes. The Exact dataset has 60,000 exomes! Such a resource could be very useful. For example, the APP variants in our exome that read out as disease causing (that is probably deterministic Alzheimer variants) in Mutation Taster were not found in Exact. This obviously makes me question the validity of the APP variants. As these exomic databases grow further in size, sequencers will be able to discard low quality or unlikely variants that do not appear in these databases which will make exomic interpretation much easier. It would be very useful if governments made it a priority to sequence their populations in order that every citizen would be within 4 or 5 generations of a sequenced genome.

Another great find was the Mutation Taster service that allows you to upload your exome vcf file onto their server for analysis. I had been entering each variant of interest singly onto their website. Uploading the entire file makes it easier though there were a lot of disease causing variants outputted.

Putting all these elements together makes a fairly powerful platform for exomic analysis. In Opal, I created an Alzheimer gene panel containing about 40 genes (e.g. from IGAP, TOP 10 from Alzforum, etc.). When I filtered out the variants above a few percent frequency, what I had left were a few dozen very plausible candidates. Many of these variants were not present in Exact, so they were of suspect validity. This approach will become even more valid when Alzheimer genetics has been more completely defined. In particular the full exome and full genome sequencing efforts now underway in AD will allow rapid analysis of AD risk of personal sequence files.

[Julie G]

Wow, J11- it seems as if you ARE making substantive progress. it's been awe inspiring to see you go through this process. Keep up the great work & keep us posted on your progress.

[J11]

I am not as sure now that I am on the right track with the exome scan analysis.

A recent article on alzforum is making me question my confidence.

Alzforum recently reported that AD genetic experts believe that only approximately 30% of AD genetic risk can currently be accounted for.
They also believe that 90-95% of familial AD genetic risk is unaccounted for. I had understood that 70-80% of the population attributable fraction of AD had already been accounted for in the GWAS studies. Thus, I thought it would be reasonable to use the existing gene set as the targets in searching for my family's risk variant. Using such a gene set approach dramatically reduces the complexity of the search. (There are currently only about 30 validated AD genes.)

Apparently this was not correct. Most of the sporadic familial AD can not be explained by genetics at this time. However, with current genetic technology it is not unreasonable to expect that large genetic risk variants should soon be easily spotted.

It is interesting that even large GWAS studies might not have been able to pick up on rare high risk AD variants, though perhaps not that unexpected.

[J11]

I am always on the look out for new Alz variants, so here is one that might be of interest.

Alzforum noted that some of the Colombian E280A PS1 mutation carriers did not develop dementia when expected. The E280A PS1 almost guarantees developing early onset AD. Genetic analysis found a protective haplotype that delayed progression by about 10 years. The SNP of interest was "rs1129844, marked eotain-1, also called CCL11". This SNP is on the 23andme genechip. I have not found any articles on this on pubmed. I am not clear which allele is risk and which is protective. I would be interested to know.

Delaying progression of AD by ten years could be very helpful.

http://www.alzforum.org/news/conference ... e-pathways

[Gilgamesh]

J11- Interesting! Have you written to Kosik and asked which is the protective allele?

[J11]

Hello again G! I hope you are enjoying the nice summer weather (writing retreats are best done in the Winter).

I have not emailed anyone about this. My assumption is that the protective allele must be the A genotype. This allele has frequency of approximately 15%.

I do find it odd that this was reported on Alzforum in April and there still appears to be no pubmed references about it. Protective alleles, such as rs1129844, would obviously be of enormous importance to those on this forum. From the Colombian experience, it was shown that people with the protective allele will still develop early onset AD, though 10 years past the expected age of onset. I am quite sure that some 44s on this forum would be very interested in knowing whether this or other alleles might be of similar benefit to them.

Alzforum (in the url I noted above) further notes that the protective haplotype with rs1129844 was confirmed to delay age of onset in another AD population (presumably more typical) of about 150 people. The Colombian mutation is carried by a large number of people; only a few seem to be protected according to the Alzforum report. It might be that only carriers of AA genotype are protected (Only 2% of people have AA genotype). However, the small size of the add on study suggests that even a single A could help. Interestingly the 23andme genotype for our loved one at rs1129844 is AG. It makes us wonder whether there is yet some large negative genetic risk factor that is lurking undetected in their genome.

It should also be noted that rs1129844 has three alleles: A, G and C (possibly also T). dbsnp reports that 26% of a European sample genotyped AG.

It is difficult to believe that such genetic protection has not already been discovered. Has the focus in GWAS been so narrowly on disease SNPs that protective SNPs have been entirely overlooked? If so, the IGAP study with 75,000 might be a good place to start the search for more such protective variants. Narrowing down the genetic risk profile for AD would be of considerable benefit when conducting AD trials as it would remove noise from the dataset.

[circular]

J11 I've just gotten around to reading this thread. Copious thanks for recording all your thoughts and experience and fine, nuanced analysis!

I looked for the new one you found, rs1129844, in Promethease but it doesn't come up. I think I am on their V4 chip. I wonder if it's only on their V3 chips?

Following ...

[Matisse]

Yes, it looks like it's only on the V3. My brother's testing was on V3 and mine and some other family members were on V4 and it says not genotyped. My brother is GG on this snp.