Discussion in 'Other Health News and Research' started by zzz, Jul 27, 2016.
Here's an inspiration for fundraising projects. From The ALS Association:
"Full" text (excluding methodology and most of the relevant data) is at http://www.nature.com.sci-hub.cc/ng/journal/vaop/ncurrent/full/ng.3626.html
Supplementary data is at http://www.nature.com/ng/journal/vaop/ncurrent/extref/ng.3626-S1.pdf
Based on reading the paper, it sounds like the results are being over-sold. First of all, the SNPs discussed are supposedly raising the risk of developing ALS - they are not causing ALS. And the odds ratio overall seems to be 2.41, which means people with one of those SNPs is only 2.41 times as likely to develop ALS. As odds ratios go, this is pretty low, and only gets higher with further processing of the data. In fact, in the US subsample, having those SNPs reduced the chance of getting ALS
I'm also a bit suspicious that they ended up keeping 74% of the patients and only 53% of the controls. One (or two?) of the filtering steps involved "out-lying relatedness" which might mean that controls were excluded due to their genetic origins being distantly related to patients. That might very well have the effect of excluding people who had the same mutations but not the same disease, if those mutations are related to genetic origins rather than diseases. At the very least, it would be good to see the rates of those SNPs in the controls before half of them were purged.
The other major problem is that this does not suggest a cure or treatment, despite what is claimed in pretty much every news article about the research. It doesn't even really suggest anything about the pathology of ALS, aside from a possible weak correlation with some SNPs.
Finally, this is published in the journal as a letter, not as a peer-reviewed research article. It's also a bit odd for a crowd-funded research initiative to have the results hidden behind a paywall
You can also try a Google Site Search
Separate names with a comma.