I am doing an overrepresentation analysis with enrichKEGG. What I am wondering, is which gene set to use in this situation:

1. I did GWAS and obtained a list of significant variants, and genes linked to them --> first set of genes
2. Then I did fine-mapping of significant variants and narrowed down the number of variants from step 1. and genes linked to them --> second set of genes
3. Then I use a cutoff of 0.5 to keep only the fine-mapped variants with a high chance of being causal and genes linked to them --> third set of genes

I am interested only in variants and genes from set 3, and I did enrichment analysis for this set, however, many genes seem not to be enriched in any pathway. I used a p-value cutoff of 0.05 and a q-value cutoff of 0.25. Should I use gene set 2 maybe? I am confused.