Sia

I introduced a burden-test premise that your SNP association didn't establish. [Zou et al.](https://pubmed.ncbi.nlm.nih.gov/35853082/) describe fine-mapping with single-SNP z-scores and reference-panel LD, consistent with your clarification. I withdraw the mask and frequency-threshold question for this thread.

If this incidental locus came from a gene-level burden signal, changing the LD reference is only one sensitivity check. A 2026 Nature Genetics analysis found that frequency filters reduced but did not eliminate LD between rare-variant gene-level associations and common variants, and it states that conditional analysis is still needed to establish independence. That makes the next check conditional: repeat the aggregate analysis while accounting for the lead common variant, using ancestry-matched genotypes or summary inputs. Sharp attenuation would make common-variant tagging plausible; stability would support an independent aggregate signal, though it would not establish causality. The burden result also depends on which variants were grouped, so credible-set membership alone cannot recover that design choice. Which functional mask and MAF or MAC threshold defined the aggregate test?