jiabowang/GAPIT

Discrepancies in GWAS Results with Updated Version

VidhiPriyanka13 opened this issue · 13 comments

VidhiPriyanka13 commented

Hello GAPIT Forum,

After updating to the latest version, I'm getting different GWAS results from the same data file. Also, there are discrepancies between analyses run on different machines. Any insights on the necessary steps for the updated version and why results differ would be appreciated.

Thank you
P

yuru0503 commented

I also found out it had different results between GAPIT3.4 and GAPIT3.3. Based on my understanding of my GWAS research, I think that the outputs from the GAPIT3.3 version are the correct ones.

Hadi-Hoda-AghBaba commented

Hi there, are you using the same model?

VidhiPriyanka13 commented
Hadi-Hoda-AghBaba commented

Could you please send both codes here.
Cheers

Hadi-Hoda-AghBaba commented

Also the Manhattan plots as well. tnx

VidhiPriyanka13 commented
Hadi-Hoda-AghBaba commented

Hi,

this link doesn;t work. Pls send a screen shot of your Manhattan plots.

Try this code to see if it solves the problem:

myGAPIT <- GAPIT
(Y=myY,
G=myG,
NJtree.group=5,
NJtree.type=c("fan"),
kinship.algorithm = "VanRaden",
model=c("FarmCPU"),
Multiple_analysis=TRUE,
Model.selection = TRUE)

VidhiPriyanka13 commented
Hadi-Hoda-AghBaba commented

Hi there,
yes good idea. Send me through the data.
Let's look at the data and we'll go from there :)
Cheers

jiabowang commented

Recently, we are revising some functions. So that makes a mistake.
Now, through comparison results and revision, the results should be correct.

Lightoscope commented

I don't know if it's related, but I'm getting different results when GAPIT calculates kinship and when that output file, 'GAPIT.Genotype.Kin_Zhang.csv' in this case, is used as KI in an otherwise identical run.

VidhiPriyanka13 commented
VidhiPriyanka13 commented