Reproducibility in urine peptidome profiling using MALDI‐TOF

MALDI‐TOF profiling of low molecular weight peptides (peptidome) usage is limited due to the lack of reproducibility from the confounding inferences of sample preparation, data acquisition, and processing. We applied MALDI‐TOF analysis to profile urine peptidome with the aims to: (i) compare centrifugal ultrafiltration and dialysis pretreatments, (ii) determine whether using signal LOD (sLOD), together with data normalization, may reduce MALDI‐TOF variability. We also investigated the influence of peaks detection on reproducibility. Dialysis allowed to obtain better MALDI‐TOF spectra than ultrafiltration. Within the 1000–4000 m/z range, we identified 120 and 129 peaks in intra‐ and interassay studies, respectively. To estimate the sLOD, serial dilution of pooled urines up to 1/256 were analyzed in triplicate. Six data normalization strategies were investigated–the mean, median, internal standard, relative intensity, TIC, and linear rescaling normalization. Normalization methods alone performed poorly in reducing features variability while when combined to sLOD adjustment showed an overall reduction in features CVs. Applying a feedback signal processing approach, after median normalization and sLOD adjustment, CVs were reduced from 103 to 26% and 113 to 25% for the intra‐ and interassay, respectively, and spectra became more comparable in terms of data dispersion.