Afanc/proteoflux: Proteoflux v1.8.5

Improved median equalization by tag Median equalization by tag is now much more robust. The normalization now uses only the intersection of features detected in all relevant tagged samples, which avoids bias from features missing in part of the selected reference/tag group. More robust Spectronaut parsing Fixed crashes caused by Spectronaut report columns that can unexpectedly contain mixed values, for example semicolon-separated entries in fields that otherwise look numeric, such as iBAQ, molecular weight, peptide position, and similar columns. Batch-effect handling Batch-effect columns are now always treated as categorical covariates, never as numerical variables. Additional validation checks were added to reduce the risk of incorrect batch-effect usage. PELSA control-point filtering Added an option to require a minimum number of measured control-concentration points for PELSA curve fitting. This helps exclude curves with insufficient control evidence. To see this option, generate a new config file. Export fixes Fixed several crash cases and formatting issues in ProteoFlux .xlsx and .pdf exports, including PELSA-related export edge cases. Other fixes Several smaller stability and cleanup fixes were included.