gusevlab / fusion_twas

methods for functional summary-based imputation
http://gusevlab.org/projects/fusion/
GNU General Public License v3.0
75 stars 45 forks source link

missing GWAS Z-scores that could not be imputed, skipping this gene. #63

Open Wanjie-Feng opened 12 months ago

Wanjie-Feng commented 12 months ago

Hi, I built my own model weights using my genotype and gene expression data, and then I had some problems running FUSION.assoc_test.R:

WARNING :   ALL/ALL.Glyma.16G000100.wgt.RDat Glyma.16G000100 16 552 4124 had missing GWAS Z-scores that could not be imputed,  skipping this gene.
WARNING :   ALL/ALL.Glyma.16G000300.wgt.RDat Glyma.16G000300 16 25166 31978 had missing GWAS Z-scores that could not be imputed,  skipping this gene.
WARNING :   ALL/ALL.Glyma.16G000600.wgt.RDat Glyma.16G000600 16 46066 46251 had missing GWAS Z-scores that could not be imputed,  skipping this gene.
WARNING :   ALL/ALL.Glyma.16G000800.wgt.RDat Glyma.16G000800 16 61778 66201 had missing GWAS Z-scores that could not be imputed,  skipping this gene.
WARNING :   ALL/ALL.Glyma.16G000900.wgt.RDat Glyma.16G000900 16 64648 64878 had missing GWAS Z-scores that could not be imputed,  skipping this gene.

A lot of genes do this, and I don't know why,my Z-score is derived from bets/se, A1and A2 are from the last two columns of the bimfile in plink format I went back to check my weights and found there were a lot of warnings

WARNING : 2 SNPs could not be scaled and were zeroed out, make sure all SNPs are polymorphic
WARNING : 3 SNPs could not be scaled and were zeroed out, make sure all SNPs are polymorphic
tmp/1_5288/twas_sample_gene.tsv.Glyma.01G019200.tmp.cv.LASSO.lasso  LASSO output did not exist
WARNING : 3 SNPs could not be scaled and were zeroed out, make sure all SNPs are polymorphic
WARNING : 5 SNPs could not be scaled and were zeroed out, make sure all SNPs are polymorphic
WARNING : 6 SNPs could not be scaled and were zeroed out, make sure all SNPs are polymorphic

Can you help me? I wonder what the problem is thank you

Best Wanjie