Proteomics of EVs from human beta-cells

The raw MS data files were processed using Max Quant software (version 1.6.7.0). Peak lists were searched against the forward and reverse Swissprot database (Homo sapiens, 20 394 sequences downloaded on February 8, 2021) using the integrated Andromeda search engine . The false discovery rate (FDR) for the peptide and protein identification was set to 1%. All proteins that cannot be distinguished based on the identified peptides were merged into one protein group. Relative quantification and normalization were performed with the MaxLFQ label-free algorithm. The MaxQuant output table was further processed with the use of the Perseus platform (version 1.6.15.0). The proteins identified in the decoy database, contaminants, and proteins only identified by site were filtered out. All bioinformatic analyses were executed on LFQ intensities transformed to a logarithmic scale with base two. The Student’s t-test with the permutation-based FDR set to 5% was used to reveal changes in protein abundances under NG and HG conditions. The statistical analysis was performed for the proteins with a minimum of 3 valid LFQ intensity values in both groups. Proteins were considered as a differential when they were identified based on at least 2 peptides and their fold change was of at least 1.5.