Subject: UKB-PPP: Request on REGENIE production flags (v2.2.1), unclipped INFO definitions, and chromosome-level variant exclusionsDear UKB-PPP Consortium / Maintainers,In our ongoing analysis using the UKB-PPP summary statistics (e.g., F2 OID30723, RSPO3 OID21415, TNFRSF11A OID20646), we require clarification on the exact production pipeline to ensure accurate downstream colocalization:REGENIE Production Details & INFO Definition:Could you confirm the exact REGENIE release build used (publication notes v2.2.1) and whether the input was BGEN v1.2 or PGEN?We observe raw floating-point INFO values > 1 in the summary statistics. Could you confirm the exact INFO calculation formula and confirm that these should remain unclipped rather than capped at 1.0?Chromosome-level Uncalled / Filtered Variants:Across full discovery chromosomes, we identified absent variant IDs (e.g., 4,021 positions in F2, 4,177 in RSPO3, 4,147 in TNFRSF11A).Could you clarify whether these variants were genuinely uncalled / filtered out during standard QC thresholds (e.g., MAC/MAF/HWE/imputation quality) in the production run, or if any separate statistical outputs exist? (We strictly maintain that genuinely unfiltered/untested variants cannot be imputed across cohorts).Matched In-study LD Matrices:If reference correlation matrices can be provided for these cis-regions, what are the exact specifications (signed Pearson $r$ vs $r^2$, alignment to EA, effective reference $N$, and phenotype normalization/residualization)?Thank you for your guidance and support.