The order of samples differs, as specified in the .fam files, between the Affy directly-genotyped and the corresponding imputed genotypes. There's a .fam file for the Illumina directly-genotyped data, but there isn't one for the corresponding imputed data set. In what order are the samples in the .chop.dosage.gz files in "ROSMAP imputed dosage chop_Illumina"?

Created by James Scherschel lilly-js
Mette? Thank you very much?I will look forward to the data be available and follow the updates discussion thread~ Best, Xiaoyan
The eQTL data is currently only available to investigators participating in the AMP-AD consortium. It will become available to other investigators in the research community at a future date. I encourage you to follow the updates discussion thread. That way you will receive an email every time new data is released. Please see here (click on the eye on the right of the page to follow): https://www.synapse.org/#!Synapse:syn2580853/discussion/threadId=2206 Best, Mette
Dear Solly, Sorry to bother you. We have written Mette a letter of application. May I ask if her email address is "Mette@synapse.org"? We haven't received any reply yet, wondering where the problem is. Xiaoyan
Thank you for your reply. I'll do as you say.
Xiaoyan- These data are not yet public. Currently, you can only access them if you're a member of AMP-AD (https://www.synapse.org/#!Team:3320424), which I don't see you in. If you are affiliated with a member institution team, your PI needs to send an email to @Mette to get you added to the Synapse group. Solly
Thanks Solly and James! Synapse said that I'm not authorized to see https://www.synapse.org/#!Synapse:syn7450279 too. I clicked the link to request access from the owner, Solly ,Could you please pass it for me?
That's done it; I can see the HRC folder now. Thank you!
@lilly-js, I think I located the problem. Please try again now.
Blek. I'm not sure what the problem might be. I was logged in. Though it seemed silly, I logged out and logged back in just in case. For self-entertainment, I tried the "show", "list", and "get" via the python command-line synapse client and received two "SynapseHTTPError: 403 Client Error: Forbidden" responses and a "No results visible to lilly-js for id syn7450279" message. Is syn7450279 perhaps accessible but, at some parent folder level, in a folder that is not? (Do you happen to know the chain of folders from there to the root folder?) Are you aware of anyone else (with only "AMP-AD Consortium" membership) successfully downloading syn7450279? I have no clue to whom the access request might have gone... I'd have to have access to see who is listed as the owner. :-/
@lilly-js, I'm not sure why you don't have access. I see that you're a member of @AMPADConsortium, which has Download access to that folder. I also didn't get a data access request for the folder. Are you sure you were logged in?
Great! Thanks, Solly. I'd been watching for the imputations to show up in the Knowledge Portal. Synapse said that I'm not authorized to see https://www.synapse.org/#!Synapse:syn7450279, so I clicked the link to request access from the owner.
@lilly-js and @adelalady - The HRC ROSMAP imputations are available to Consortium members with ROSMAP access: https://www.synapse.org/#!Synapse:syn7450279. I've also generated eQTL for these if you're interested. This is the most recent version with the current RNA-seq normalization: https://www.synapse.org/#!Synapse:syn9884468.
Ben/Solly: Any idea when the HRC-based imputation will be uploaded? Have they already been somewhere? With eQTL data? Xiaoyan Li: When I last looked at these data ~7mo ago, I encountered a few non-trivial issues. I'd strongly recommend basing any analyses you might have planned on the newly imputed, via HRC, genotypes if possible. As alluded to earlier in this discussion, the directly-measured Illumina genotypes were GRCh37 while the Affy genotypes were GRCh36. If I recall correctly, the imputed genotypes appeared to have been imputed against GRCh37, presumably after a lift-over of the GRCh36 Affy genotypes, for both platforms... (My recollection is hazy on this particular point.) The Illumina imputation was highly concordant with genotypes derived from the RNASeq data (lifted from GRCh38 back to GRCh37 to put all the data on the same genome build), but the Affy imputation was highly DIScordant. You'd probably want to make sure that the Illumina and Affy genotypes were on the same genome build (lifting over as required), re-impute, QC the results, and verify appropriate levels of concordance amongst the (DNA array and RNASeq) data sets. You would, as part of that, end up testing the consistency of ROSMAP_IDkey.csv (syn3382527) with what you saw in the genetic data. You should expect to encounter issues there as well and, likely, would end up creating your own cross-dataset key mapping consistent with the results of that work. All of that seems like a lot of duplicated effort if the HRC-based imputation results are available... Ben/Solly: Any idea when the HRC-based imputation will be uploaded? Have they already been somewhere? With eQTL data?
There's a .fam file for the Illumina directly-genotyped data, but there isn't one for the corresponding imputed data set. Are there have the .fam file in "ROSMAP imputed dosage chop_Illumina?
That works. It's easy enough to map the b37 imputed genotypes to b38 after the fact.
@lilly-js, All of the samples have been imputed to the HRC reference, which is the most recent imputation panel. It uses GRCh37 (hg19). To my knowledge, there is no current imputation panel to build 38. All variants in the panel are included in the output, including those that have been assayed by the genotype panel.
Thanks, Ben. Both the Affy- and Illumina-genotyped samples? Will all of that be on genome build 38? (Instead of the current mix of b36 and b37?) Presuming that the option is/was available, are the directly-genotyped variants included in the imputation output? It's a nice sanity-check when they are...
It looks like that information is truly missing. We re-imputed the genotypes, and will be providing that data soon. Thanks, Ben

.fam missing for "ROSMAP imputed dosage chop_Illumina" (imputed genotypes in folder syn2426141) page is loading…