search

JonJala / mtag

Python command line tool for Multi-Trait Analysis of GWAS (MTAG)
GNU General Public License v3.0
171 stars 54 forks source link

gwas lambda inflation factor #92

Open kys21207 opened 4 years ago

kys21207 commented 4 years ago

Hi there,

Recently, I combined 4 different traits using MTAG without any error or warning. However, I concern about the lambda inflation factor because the lambda is higher than expected. Please see the examples,

1 trait: gwas(1.0475) vs mtag(1.0765) 2 trait: gwas(1.0475) vs. mtag(1.1543) 3 trait: gwas(1.0475) vs. mtag (1.1751) 4 triat: gwas(1.0966) vs. mtag(1.2002) What do you think?

Second question is I lost so many variants (75%) due to strand ambiguity. gwas (38,500,986) vs. mtag(9445190). Our gwas data came from UKBiobank.. I am not sure why there are so many strand ambiguous variants. Any suggestion to check??

Thank you.. KIjoung

paturley commented 4 years ago

Hello Kijoung,

I'm not totally surprised by the results you showed me, but it's hard to say without knowing details about the phenotypes you are studying. In general, lambda_GC is increasing with increased statistical power and MTAG should increase statistical power.

I'm also not surprised by the loss of strand ambiguous SNPs, but it's worth verifying that the SNPs that are being dropped are in fact strand ambiguous. I'd look up a handful and see.

Patrick

On Mon, May 4, 2020 at 3:06 PM Kijoung Song notifications@github.com wrote:

Hi there,

Recently, I combined 4 different traits using MTAG without any error or warning. However, I concern about the lambda inflation factor because the lambda is higher than expected. Please see the examples,

1 trait: gwas(1.0475) vs mtag(1.0765) 2 trait: gwas(1.0475) vs. mtag(1.1543) 3 trait: gwas(1.0475) vs. mtag (1.1751) 4 triat: gwas(1.0966) vs. mtag(1.2002) What do you think?

Second question is I lost so many variants (75%) due to strand ambiguity. gwas (38,500,986) vs. mtag(9445190). Our gwas data came from UKBiobank.. I am not sure why there are so many strand ambiguous variants. Any suggestion to check??

Thank you.. KIjoung

— You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub https://github.com/omeed-maghzian/mtag/issues/92, or unsubscribe https://github.com/notifications/unsubscribe-auth/AFBUB5OWAMT64VMJRQG6YSLRP4G2RANCNFSM4MZAGCCQ .