A comparison of four proteomics software for hair proteome analyses

1.1Hair has applications in biomarker discovery and forensics, yet the influence of proteomics software tools on hair proteome characterisation remains underexplored. This study compares four bottom-up proteomics workflows (MaxQuant, FragPipe, MetaMorpheus, and SearchGUI/PeptideShaker). Publicly available hair proteomes were analysed following extraction with 1-dodecyl-3-methylimidazolium chloride (DMC), sodium dodecanoate (SDD), sodium dodecyl sulfate (SDS), and urea. Data were acquired on Orbitrap-based DDA platforms. Peptide identification, protein inference, functional annotation, physicochemical properties, and label-free quantification (LFQ) were evaluated. Peptide-level performance differed across tools. MS-GF+ and FragPipe identified the most unique peptides, while X!Tandem reported the fewest. Protein inference showed a dissociation from peptide-level results. MetaMorpheus reported the highest number of protein groups despite only the third highest peptide counts. FragPipe and MaxQuant followed, while PeptideShaker consistently inferred the fewest proteins. Protein-level concordance was low, with only 30.3% overlap across tools and extraction methods. These differences extended to downstream analyses. Functional enrichment showed moderate concordance (38.25% overlap). Physicochemical profiles varied, with MetaMorpheus identifying more hydrophobic proteomes and PeptideShaker more hydrophilic profiles. At the quantitative level, reproducibility depended on extraction buffer. SDS and urea showed lower variability (CV =< 0.025), while DMC and SDD showed higher variability (up to 0.10). Absolute LFQ intensities and differential expression outputs varied across tools despite moderate to strong correlation (r = 0.77 to 0.93). Overall, software choice influences proteome coverage, physicochemical profiles, and quantitative outcomes. Relative trends were partially conserved, but magnitude and significance varied. These findings support careful method selection and multi-tool validation in hair proteomics