Phenotype for Plink GWAS data

146 views
Skip to first unread message

Nalani Knight

unread,
Apr 19, 2020, 10:53:42 PM4/19/20
to Alzheimer's Disease Neuroimaging Initiative (ADNI) Data
Hi all,

I am new to PLINK and running GWAS, however I've found the online resources for PLINK very helpful. I've been able to download the ADNI data file for PLINK and have done some basic analyses. However, the phenotype column in this data is missing. Does anyone know if ADNI provides a GWAS file that includes an indicator of AD diagnosis (the GWAS ADNI1 data file from the database does not have this info - I've checked)? This would be extremely helpful for performing an association analysis in PLINK. Or, would someone perhaps more experienced in PLINK be able to provide me with code or a resource to help with this problem? 

Thank you all for your help. Stay healthy!

Nalani

Danielle J Harvey

unread,
Apr 20, 2020, 1:38:20 PM4/20/20
to adni...@googlegroups.com

Hi Nalani,

 

I haven’t worked with the GWAS data from ADNI, but there is information on diagnosis on the Study Data page. The Diagnostic Summary table, for example, contains the diagnosis at each visit for each individual (RID). You can refer to the ADNI data dictionary (also available on the Study Data page) for the coding of the DXCURREN variable.

 

I just downloaded some of the .csv files (rather than the PLINK) for the ADNI-1 GWAS data. It looks like Sample ID is in the format XXX_S_YYYY (the numbers in YYYY correspond to the RID).

 

Not sure if this is helpful or not…

 

Danielle

--
You received this message because you are subscribed to the Google Groups "Alzheimer's Disease Neuroimaging Initiative (ADNI) Data" group.
To unsubscribe from this group and stop receiving emails from it, send an email to adni-data+...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/adni-data/1edb6d0e-7dfa-401a-b90e-ad4a64160dd0%40googlegroups.com.

Nalani Knight

unread,
Apr 22, 2020, 10:44:03 AM4/22/20
to Alzheimer's Disease Neuroimaging Initiative (ADNI) Data
Hi Danielle,

Thank you! This is helpful. I was using some of the imaging data from ADNIMERGE to merge by the RID variable (after converting from PTID), but I see that using DXCURREN will help even more. 

I figured out how to add a phenotype to the PLINK data. Turns out it is possible to run PLINK from R, so I can read in the dataframe from PLINK, merge it with a dataset/column that provides the phenotype and then upload it back to PLINK to run the GWAS. I love it when things work out so nicely like that. 

Have a great day!

Nalani

To unsubscribe from this group and stop receiving emails from it, send an email to adni...@googlegroups.com.

Danielle J Harvey

unread,
Apr 22, 2020, 1:19:20 PM4/22/20
to adni...@googlegroups.com

Hi Nalani,

 

Great! ADNIMERGE also has the diagnosis information in it (DX_bl is the screening diagnosis while DX is the diagnosis at different visits).

To unsubscribe from this group and stop receiving emails from it, send an email to adni-data+...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/adni-data/a278e78b-c1bf-4ac6-acd8-7e399009b170%40googlegroups.com.

Nalani Knight

unread,
Apr 22, 2020, 1:39:05 PM4/22/20
to adni...@googlegroups.com
Danielle,

Yes I’ve noticed this. I also noticed that DX_bl has five categories (CN, SMC, EMCI, LMCI and AD) and  DX has only 3 (CN, MCI, AD) as does DXCURREN. This makes it challenging to predict a conversion into the subsequent diagnosis categories at later time points because we don’t know whether the MCI stands for either of the 3 mid-categories (SMC, EMCI or LMCI). There’s probably another data set that further breaks this up into the 5 categories. Do you happen to know if this is true offhand? Otherwise I’ll keep digging! 

Thanks!

Nalani 

Danielle J Harvey

unread,
Apr 22, 2020, 1:46:39 PM4/22/20
to adni...@googlegroups.com

Hi Nalani,

 

Just some background for you: In ADNI-1 (RID<2000), we only had 3 diagnostic groups (NL, MCI, AD), but based on our definition for MCI, they were really deemed to be “later” in MCI (based on a variety of data, this group was much closer to the AD group than the NL group) – we called them LMCI. So, in ADNI-GO (2000<RID<4000), we newly enrolled folks earlier in the course of MCI (EMCI). In ADNI-2 (4000<RID<6000), we had NL, MCI, and AD, but MCI included both EMCI and LMCI folks. We also added SMC (subjective memory complaint/concern) – these folks are NL, but have a concern about their memory.

 

According to the Clinical Core, the diagnoses are really NL, MCI, and AD, so from their perspective, knowing whether a participant with MCI is EMCI or LMCI or progresses from EMCI to LMCI isn’t all that meaningful, since they are all MCI. That’s why the information in the Diagnostic Summary and ADNIMERGE (for the diagnosis at each visit) only includes the 3 categories.

Michael Romano

unread,
Apr 22, 2020, 1:51:11 PM4/22/20
to Alzheimer's Disease Neuroimaging Initiative (ADNI) Data
Thanks!

Also following along. I've been using these codes from the DXSUM sheet. SMC falls into NL? Also are CN and NL the same?

Best,
Mike

--
You received this message because you are subscribed to the Google Groups "Alzheimer's Disease Neuroimaging Initiative (ADNI) Data" group.

To unsubscribe from this group and stop receiving emails from it, send an email to adni...@googlegroups.com.

--
You received this message because you are subscribed to the Google Groups "Alzheimer's Disease Neuroimaging Initiative (ADNI) Data" group.

To unsubscribe from this group and stop receiving emails from it, send an email to adni...@googlegroups.com.

Danielle J Harvey

unread,
Apr 22, 2020, 1:51:56 PM4/22/20
to adni...@googlegroups.com

Yes, SMC falls into NL and CN and NL are the same.

 

Danielle

To unsubscribe from this group and stop receiving emails from it, send an email to adni-data+...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/adni-data/ab4ce136-a9f0-4826-8696-c7f224a33e4f%40googlegroups.com.

Pav Kalinowski

unread,
Jul 15, 2021, 7:54:39 PM7/15/21
to Alzheimer's Disease Neuroimaging Initiative (ADNI) Data
Which online resources did you find useful? I'm struggling to access the plink files. I am using R and so far no success
Reply all
Reply to author
Forward
0 new messages