Big Data App to Explore Genomes for Clinical Relevance, Rare Variants, Drug Response, etc (Free)

[Rufous McKinney]

My story was: breast fed, but foods added starting at two months. At year one: Cows Milk. RASH

It was all downhill from there. (except for numerous uphill climbs, and welcomed level playing fields and mesas).

 

[kday]

So you are allergic to milk?

 

[Rufous McKinney]

in 1954: told alergic to milk. and lots of other things.

i recall years of Hating Goats Milk.

then: I don't know. I ate cheese. I just didh't exhibit those symptoms: frankly: I somewhat thought that morphed into: these types of syptoms (ME/CFS).

I've never done the restriction diet....

there is a Cow Protein allergy. maybe this is a different allele. And I don't know if I have it.

 

[Rufous McKinney]

@kday

I will read more on your design and see if I can figure it out....now I seem homozygous for something that I should be checked out by now....but its under Drugs and its a disease so I don't understand the DRug part: so obvoiusly there is alot to understand here!!

 

[kday]

I think one thing that is probably a big risk factor for some people with ME/CFS is MBL2 deficiency.

If one shows up with two MBL2 heterozygous (or homozygous) SNPs, they are likely to have Mannose-Binding Lectin Deficiency.

My favorite genetic disease. Because 5-10% of the population has it. ME/CFS patients are about double that number. And much higher in sub-Saharan Africans.

It's like the orphan disease that everyone ignores and nobody has heard of because it's too common to pay attention to. Big risk factor for so many immune related things.

Easily one of my favorite genes/diseases (after lactase persistence of course).

I have a similar life experience as you when it comes to milk allergy. I'm fine with it now. Allergic as a kid.

 

[kday]

@Rufous McKinney
Yeah, the drug part isn't straight forward. I'm trying to rethink this section as it's not even clear to me when I look at it. Sometimes common variants are the risks. And they are generally been modified to be oriented in the direction towards the drug risk. But sometimes there are different risks for homozygous or normal variants. And sometimes a heterozygous variant is the risk. So interpret this section with much caution please. I probably have no choice but to curate this section more than I would like to.

Clicking on SNPedia links for drug risks can be useful in the meantime.

 

[rel8ted]

I can totally see why you are overwhelmed. I can tell you that it is not as complicated as it looks,

I think that is true, but I have a harder time making things stick than I used to & wish there was something to bring me up to speed in this area. I generally really enjoy technical science things.

 

[kday]

@rel8ted

I honestly couldn't get ANYTHING done for a number of years because of my brain. My brain and my heart are probably the two organs that were hit the hardest. I can't tolerate driving much for example, but trying to rehabilitate.

I found something that helped in a systemic way last year. And I have been more capable of doing complex cognitive tasks that I thought I could never do again. That said, I still get overstimulated and get autism like fits from it. Not that I have autism. More like adult-onset autism if you want to classify it that way.

 

[wigglethemouse]

I missed the saving instructions at the end of the post (I don't do well with lots of text)

So to save, just use your browser "File save page as" and the data will be stored to your local computer as a local html file for later review.

 

[kday]

I missed the saving instructions at the end of the post (I don't do well with lots of text)

So to save, just use your browser "File save page as" and the data will be stored to your local computer as a local html file for later review.

Yes, this generally works. An issue is that some font icons may be missing. I'm seeing if I can resolve this somehow. But it mostly works.

 

[wigglethemouse]

My favorite genetic disease. Because 5-10% of the population has it. ME/CFS patients are about double that number.

I'm enjoying playing with your tool. I am heterozygous to this. Where did you find out about this variation in ME/CFS patients being double the general population occurrence?

 

[kday]

I'm enjoying playing with your tool. I am heterozygous to this. Where did you find out about this variation in ME/CFS patients being double the general population occurrence?

There is a study or two. Everyone forgot about it. Just like everyone forgot about what Mannose-Binding Lectin Deficiency is.

You may or may not have a another SNP in your whole genome. I believe 23andMe misses one important variant. It might depend on the chip they used for your file.

If you don't have two heterozygous variants, you don't usually have the condition. However this is not always the case. Directly testing MBL function is a test available at any major lab. It's the only way to know for sure.

 

[kday]

@wigglethemouse

Deficiency of MBL was found in 15% of the CFS patients in contrast to 6% in a historical control group.

I was wrong, sorry, it's more than double.

https://www.ncbi.nlm.nih.gov/pubmed/26429318

 

[kday]

Rolled out some minor updates and added the most important feature: Emojis during upload. I was told it is important to not be too serious, hence why it's an important feature.

To enjoy the full experience, you have to upload your genome.

And for those who wonder how uploading their data helps. I can see graphs of CPU activity from instances (a group of many servers that automatically grows and shrinks depending on how much usage there is). I can see that the instances are scaling, and this has allowed 2-3 uploads at a time in some cases without any slow downs in speed. Seeing these charts let me see how well the app can scale and helps me tweak settings. And so far, so good. No crashes! Thanks everyone for your help!

 

[Moof]

Thank you for this great tool, I uploaded my data last night. I knew about some of the issues (MTHFR deficiency, carrier status for haemochromatosis, heterozygous for AMPD1 deficiency, red hair variant, alleged lactose tolerance) but the BRCA2 mutation came as a bit of a surprise! These mutations are quite complex, though, and the result is only based on 23andMe data, so I'm not going to worry about it unduly.

I found the tool useful and straightforward, and the interface friendly. I saved it as a web archive without any difficulty, and I'm able to open it and access all the links from my local drive. Thanks again for developing it, @kday, it must have been a huge amount of work!

 

[Rufous McKinney]

Everyone forgot about it. Just like everyone forgot about

I think there is a whole lot more of: Everybody forgot about it...than meets the eye.

 

[Rufous McKinney]

To enjoy the full experience, you have to upload your genome

Are you saying we should upload our data again to get an updated version?

 

[kday]

Are you saying we should upload our data again to get an updated version?

It was more of a joke about the Emoji's. No significant changes in terms of what you see. So, no, there will be no benefit on uploading again other than to see Emoji's. The minor fixes are to some static content in the headings that almost nobody cares about.

 

[kday]

Thank you for this great tool, I uploaded my data last night. I knew about some of the issues (MTHFR deficiency, carrier status for haemochromatosis, heterozygous for AMPD1 deficiency, red hair variant, alleged lactose tolerance) but the BRCA2 mutation came as a bit of a surprise! These mutations are quite complex, though, and the result is only based on 23andMe data, so I'm not going to worry about it unduly.

I found the tool useful and straightforward, and the interface friendly. I saved it as a web archive without any difficulty, and I'm able to open it and access all the links from my local drive. Thanks again for developing it, @kday, it must have been a huge amount of work!

If the BRCA variant says benign, it's probably no big deal. Some variants may or may not be misclassified. You can only know by checking the research. I would also check OpenSNP to see how common it is in other people. There is a link in the app you can click on.

And as stated in the first post, there is a limitation in 23andMe regarding custom identifiers (which are most BRCA variants). If your genotype is a contains a D or an I, the algorithm sort of uses a best guess approach if there are multiple variants at the same chromosomal position. So it can get some of these variants wrong. But even if it gets the exact variant wrong, there can be a similar variant that may or may not the same consequence at the same position. (Not trying to make your head spin. It's complicated.)

However, if the variant contains a D/I and the Reference Allele (in the left column) is not showing as D or I, this is a bug I need to know about.

That all said, you are right. The best way to evaluate BRCA is not through 23andMe. AncestryDNA data is more reliable if you have that. Unfortunately, 23andMe data scientists got a lot of things wrong. Bad data was removed from all chips to the best of my ability by manually cross-referencing the frequency OpenSNP for all these SNPs/indels. OpenSNP frequency is based on other users that upload genomes through services like 23andMe and AncestryDNA.

 

[kday]

Updates and bug fixes: 23andMe proprietary SNPs and indels have been converted to proper rsid. It can even differentiate between ids at the same exact chromosomal location whether it's a SNP, indel, etc. I believe this is thought to be a limitation with the 23andMe platform, but I was able to successfully engineer a workaround with high accuracy for this limitation. This is a total of about 17,000 reported ClinVar SNPs.

There is no more "best guessing" what the variant is if there are multiple at the same chromosomal location. It now knows how to translate/call the exact non-proprietary identifiers.

This may or may not change results in all. But it may get rid of some weirdness with genes like BRCA and some others.

I also made some minor changes/bug fixes to the auto-annotation.

Multi-allelic variants are not yet fully supported (this will take some time). I still need to fix the "bug" where only one SNV/indel per chromosomal location is reported. It's a simple database issue, but sifting through the data on both platforms to ensure accuracy of reporting may take some time. This bug probably won't have much impact on most results, but it can potentially miss something important in a some.

For those who had potentially erroneous BRCA variants or others, I suggest re-uploading files. This proprietary identifier issue was a 23andMe only issue.