Can you use a polygenic risk score for a trait (e.g., BMI) in association analyses when it doesn't associate with BMI in the sample it is used?

Question

[kl]

I have a polygenic risk score and after merging it with the data, I have about 500 observations. I am performing this analysis to check my my genetic data cleaning and imputation went well. However, contrary to what I expected, my polygenic risk score for BMI does not associate with BMI in my cohort.

1. I am worried about the imputation of my genetic data (it is weird that it doesn't associate)
2. Can you use a polygenic risk score for the trait based on weights from external GWAS, when it doesn't associate strongly with the trait in your cohort?

Thanks