search

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

Using FUSION for Individual-level predictors #55

Closed jianvhuang closed 1 year ago

jianvhuang commented 1 year ago

I plan to calculate a gene expression risk score using the method described in the paper below. Oliver Pain, et al, Imputed gene expression risk scores: a functionally informed component of polygenic risk, Human Molecular Genetics, Volume 30, Issue 8, 15 April 2021.

I have used FUSION for the summary level TWAS. I want to combine the eQTL weight with my individual-level genotype data. I tried Rscript make_score.R [wgt.RDat file] > [SCORE_FILE] but sometimes the output is a dataset with all NAs of 4 columns, and sometimes the output is an empty file. In some cases, I did obtain output with SNP, A1, A2, Z.

The wgt.RDat file was downloaded from the FUSION tutorial webpage (GTEx8, whole blood and brain-related tissues). I wonder why the score files are not generated correctly for some genes. Thank you.

jianvhuang commented 1 year ago

This is most likely because I used the nofilter.pos file, instead of the .pos file which only contains genes with significant heritability

alyssacl commented 8 months ago

This is most likely because I used the nofilter.pos file, instead of the .pos file which only contains genes with significant heritability

Is it recommended to use .pos or nofilter.pos?

The answer is: Using genes that achieved significant heritability is recommended for typical analyses. http://gusevlab.org/projects/fusion/