/sci/ - Science & Math

More than bayesian, and I'm sure the test results will be explained. You don't just get a number tossed at you. The technology is certainly there in detection. And read the wiki article for BRCA.

>>5812625
>More than bayesian
Out of curiosity, do you know the algorithms implemented, or a good resource for learning about these methods?
>and I'm sure the test results will be explained.
I've heard that's not exactly true with 23andMe, and that many people do not know how to interpret the results. I think I'll be okay, though.
>You don't just get a number tossed at you. The technology is certainly there in detection. And read the wiki article for BRCA.
I'm confident that BRCA mutations will be detected, I'm not worried about that. It's whether the ancestry package and the other information (likely traits, risk factors for diseases) are worth shelling out $99 when I could get BRCA testing, alone, done cheaper.

>>5812660
>good resource, algorithms
If you're in bioinformatics, focus a bit on the genetics side of it to learn about this. As for algorithms, most of them will be heuristic models from detecting a few SNPs you have to give probabilities rather than building an entire database of the SNPs.
>worth $99 for bigger package
If you have the money, why not. It will not tell you much though that you couldn't already gather from talking to your family/grand-family about their parents - ie; 'higher' risk of heart disease, obesity, cholesterol, etc. I'm not sure what genes they specifically test but it doesn't look like any of the results will be life changing and crucial that you already aren't aware of (like BRCA).

>>5812660

>worth shelling out $99

defiantly have a Jew here

>>5812619
I did it and I love it OP.

Get ready to find out you have some black in you, because I'm whiter than whitebread despite what my genetics might lead one to believe.

>>5812709
Also,if you're into the biology they give you full access to your genome

Don't waste the money. Those tests aren't very reliable. There was a study a while back where someone submitted the same sample to a bunch of different genotyping companies and got surprisingly different results from each one.

Also, most of the information they tell you is stuff you could infer just looking in a mirror.

Also, protip: they just stick your sample on a microarray. More expensive the test - > larger the array, more probes they can do.

>>5812713
That's not your full genome. All they're doing is giving you a reference genome with any SNPs they detected, changed in the sequence.

Even the very best technology we have hasn't pushed full genome sequencing below $1000 yet. It's a microarray.

>>5812718
That's why I said "full access" not "full genome". Thanks for your input though!

>>5812717
>Also, most of the information they tell you is stuff you could infer just looking in a mirror.
I must be pretty good if I can infer my increased risk of coronary heart disease by looking into a mirror.

>>5812724
>Coronary artery disease has a number of well determined risk factors. The most common risk factors include smoking, family history, hypertension, obesity, diabetes, high alcohol consumption, lack of exercise, stress, and hyperlipidemia. Smoking is associated with about 54% of cases and obesity 20%. Lack of exercise has been linked to 7–12% of cases.
You probably could actually infer increased risk of coronary heart disease by looking in the mirror.

>>5812730
God I hate people like you. You're that pompous prick in class with no friends who "secretly" thinks he's better than everyone else.

>>5812737
No, I'm the sadsack with no friends who has crippingly low self-esteem.

>>5812740
sadsack.jpg?

>>5812743

>>5812687
>If you're in bioinformatics, focus a bit on the genetics side of it to learn about this. As for algorithms, most of them will be heuristic models from detecting a few SNPs you have to give probabilities rather than building an entire database of the SNPs.
My background is actually in EE. I'm studying neural applications of DSP.

>>5812703
So you would shell out $99 for a test that you don't know provides information of use?

If you really want your genome sequenced, buy the reagents for making a multiplexed library and prep it yourself.

Find a lab that's doing some sequencing experiments now and bribe somebody with a couple cases of beer to put your sample in with theirs.

Geneticist here, what do you want to know?

>>5812718
>Even the very best technology we have hasn't pushed full genome sequencing below $1000 yet. It's a microarray.
How good is the information I'd get with full genome sequencing? Could I possibly I get the data on file in case newer information is discovered and I'd want to try running detection algorithms on it?

>>5812717
>Don't waste the money. Those tests aren't very reliable. There was a study a while back where someone submitted the same sample to a bunch of different genotyping companies and got surprisingly different results from each one.
Do you attribute this to the technology not being there yet? "A while back" can be very significant in biotechnology.

>>5812768
How good the information is with a service doing SNP, both for medical information (some risk factors clearly aren't as useful of information as, say, learning you're a character for an awful disease) and for ancestry. I'm pretty sure it's just SNP tests, no full sequencing.

>>5812785
>How good is the information I'd get with full genome sequencing?
With modern sequencing technology, very good.

>Could I possibly I get the data on file in case newer information is discovered and I'd want to try running detection algorithms on it?
Probably? Find out before you drop the money, but I'd assume the answer is yes.

>Do you attribute this to the technology not being there yet?
It's not so much that the tech isn't there yet so much as the tech they were using has some inherent limitations. Microarrays were in style for a good decade because they were dirt cheap, but they really do have their issues that can make analysis more of a one-off effort than something truly replicable.

>>5812802
So, that said, is 23andMe worth it if I lack the cash for full sequencing?
Which companies, in your experiences, give the best deal for the information you get?

bumpin4interest

>see this thread about 23 and me
>has 23 posts in it
>oh how coincidental
>phone also has 23% charge at the same time
>cum buckets

>>5813852
the nummers maaan
what (or when?) do they meeaaan

>>5813869
they mean that i just experienced something that's probabilistically very rare.
should've used up my one lucky moment in life on a lottery ticket, fuck.

>>5813873
Totally how probability works....