Fig. 4

Perseus software was used to perform statistical analysis on the MS data. A Heatmap of unsupervised clustering where normalisation was based on z scores and the data for each row was normalised by the mean value calculated for the same row. Green represents low protein intensity and red representing high protein intensity. B Principal component analysis was performed. C Venn diagram was created to represent the total proteins identified in the two cell variants’-derived EVs and the n = 5 proteins that are specific to each. D The total vesiclepedia database (as of 12 July 2022) was compared with the proteins identified. Cellular (E) and biological (F) components significantly associated with the proteins identified in TNBC EVs. All analysis was performed using n = 3 biological repeats