Where communities thrive


  • Join over 1.5M+ people
  • Join over 100K+ communities
  • Free without limits
  • Create your own community
People
Repo info
Activity
  • Aug 20 09:57

    tsujigiri on update-nokogiri

    (compare)

  • Aug 20 09:53

    gedankenstuecke on master

    Update nokogiri (#518) (compare)

  • Aug 20 09:53
    gedankenstuecke closed #518
  • Aug 20 09:53
    gedankenstuecke commented #518
  • Aug 20 09:13
    tsujigiri opened #518
  • Aug 20 09:12

    tsujigiri on update-nokogiri

    Update nokogiri (compare)

  • Jul 27 10:14

    gedankenstuecke on master

    Add disclaimer to new message e… (compare)

  • Jul 27 10:14
    gedankenstuecke closed #517
  • Jul 27 08:26
    philippbayer opened #517
  • Dec 06 2018 09:32

    tsujigiri on update-rails

    (compare)

  • Dec 06 2018 09:32
    tsujigiri commented #516
  • Dec 06 2018 09:32
    philippbayer commented #516
  • Dec 06 2018 09:32

    philippbayer on master

    Update rails (CVE-2018-16476) Fix test Remove require and 2 more (compare)

  • Dec 06 2018 09:32
    philippbayer closed #516
  • Dec 06 2018 09:30
    Travis openSNP/snpr (update-rails) fixed (2267)
  • Dec 06 2018 09:29
    Travis openSNP/snpr (update-rails) still failing (2265)
  • Dec 06 2018 09:29
    Travis openSNP/snpr#516 broken (2266)
  • Dec 06 2018 09:28
    tsujigiri synchronize #516
  • Dec 06 2018 09:28

    tsujigiri on update-rails

    Put back shoulda-context (compare)

  • Dec 06 2018 09:27
    tsujigiri synchronize #516
Bastian Greshake Tzovaras
@gedankenstuecke
Ah, that makes sense. For the generation of dumps: we only create the phenotype specific ones on request. The overall dump is already automated to be run regularly (though it seems to not work these days :p)
Philipp Bayer
@philippbayer
thanks for fixing it @gedankenstuecke @tsujigiri, well done :) I don't see an easy way to prevent this in the future except having a cron job which deletes 'old' backups?
Bastian Greshake Tzovaras
@gedankenstuecke
Yeah, I think that’s a good way to go in any case :)
Helge Rausch
@tsujigiri
Already on it.
Bastian Greshake Tzovaras
@gedankenstuecke
awesome, thanks!
Bastian Greshake Tzovaras
@gedankenstuecke
whops our SSL expired again, just generated us some new letsencrypt :)
Philipp Bayer
@philippbayer
thanks!
Mark Glasgow
@glasgowm148
Hey - any tips on tools to use to compare open SNP data ?
I want to compare the RCCX region on chromosome 6 with those diagnosed hEDS
I've done 5 manually and the results are already what I expected. Out of 10 mutations, One was repeated everytime (normal frequency 28% - 38% on opensnp), 2 were repeated 60% of the time (One normal freq 7%, opensnp 11%), (one normal freq 0.04%, openSNP (probably miscall due to region), and 1 was repeated 40% of the time (Normal frequency 2%, opensnp 4%)
I figure the slightly elevated levels are the tendency for those with chronic illness to research DNA
Bastian Greshake Tzovaras
@gedankenstuecke
Hey, cool to see that you’re looking into the data! To which reference data do you want to compare it?
Mark Glasgow
@glasgowm148
each other? or my own
To see which yet undefined mutations in RCCX are common in hEDS + chronic illness (POTS/MCAS/FM/CFS/autoimmune)
as I believe they are all the same condition
Bastian Greshake Tzovaras
@gedankenstuecke
and you have a set of data sets of people with those different conditions and want to compare those to all the data in openSNP to see how frequent the different variants are?
Mark Glasgow
@glasgowm148
I could get datasets for people who likely to meet RCCX criteria (multiple diagnosis) - just want to compare all those datasets with each other. So I can see what mutations occur in 'RCCXrs' at a much higher frequency than the normal population
But I'm not sure if any chip other than 23andme v4 has attempted to sequence it so a bit tedious going through each file by hand
Bastian Greshake Tzovaras
@gedankenstuecke
ok, so what you’d need from openSNP is a list of frequencies for all variants the RCCX region?
Mark Glasgow
@glasgowm148
but if I could load 300 datasets (of people with hEDS + ADHD, for example) into some sort of software - and it tells me what RCCX mutations are most common between those 300 - could see some interesting things
Basically to show a correlation between these conditions and RCCX mutations? I figure that must be a thing that's used frequently but not having much luck finding the right tool
Bastian Greshake Tzovaras
@gedankenstuecke
so i think what you would need to do is compare the frequency of variants in your population of e.g. people with hEDS + ADHD to the frequency in a larger general population and see how they differ?
Mark Glasgow
@glasgowm148
sure - but don't we already know the frequency in the general population ?
I'm just looking for a tool that can do the first bit :P
Bastian Greshake Tzovaras
@gedankenstuecke
i guess my Q is whether you already have this population of people with hEDS + ADHD or is finding that population the issue? :D
Mark Glasgow
@glasgowm148
Nope, I can use educated guesses on opensnp (although the categories are a bit crap) to gather data
so I can get the dataset
but what data tool would you use to compare
suppose I could figure out how to do it in R
think this does what I need :: https://github.com/HajkD/orthologr
Bastian Greshake Tzovaras
@gedankenstuecke
ah, i don’t think you want to do anything with orthologs. What you rather would need to do is running statistical tests for comparing whether the changes in frequency are significant or not. I think @philippbayer has done some work like that. maybe he has an idea for how to do it
Mark Glasgow
@glasgowm148
maybe VCFtools? I'm surprised it's not available in an executable app though. Figured none-cs people would be daft for comparing files with their relatives
Philipp Bayer
@philippbayer
isn't that an FSt test between the 'healthy' and the 'afflicted' population groups, using vcftools?
you could also make for each snp a contingency table:
              healthy    sick
alleleA    1233         5
alleleB       10        1233
and then run a million chi squares, one for each SNP, and then use something like p.adjust() to get FDR adjusted p-values
Philipp Bayer
@philippbayer
if it's a bunch of SNPs right next to each other, running something like FarmCPU via rMVP should also work
or something like https://www.pnas.org/content/116/4/1195
Mark Glasgow
@glasgowm148
Hmm, I'll have a look into that thanks. Maybe I'm misunderstanding the terminology but I don't care about the healthy group
Just want frequency in the dataset provided
Theres only a small amount of 'risk' genes I'm looking for so doing that bit manually is easier
I'm making a auto-updating spreadsheet people can select their genes from too : https://docs.google.com/spreadsheets/d/1g7f-0DT2_XYcTdjzYwE0V30LenynAL05zsQJkYXqITM/edit?usp=sharing
Bastian Greshake Tzovaras
@gedankenstuecke
The thing is that you need to compare the frequencies of the non-healthy group to something in order to know whether the change in frequency is meaningful. That’s why you need to have this one too
Mark Glasgow
@glasgowm148
I believe this affects 30-50% of the population and such comparisons are why it's missed
30% have adrenal issues/mutations (CAH) - 60% have methylation issues (MTHFR/etc) - chronic illness is a Venn diagram of those two sets
  • C4 / CYP mutations
  • +
Mark Glasgow
@glasgowm148
Ok, think I've figured it out. I think you're right in the end - I need to compare manifested diseases in those with RCCX mutations to those without and see if immunological diseases appear above expected
Is there an easy way to export everyone who has a specific mutation?
Philipp Bayer
@philippbayer
sadly not right now! you can export all specific phenotypes, but not specific snps
you could download the entire dataset and use grep to find the right files