Andrea Petretto

@angrist.bsky.social

956 Followers 1111 Following

Head of the Core Facility for Omics Sciences at IRCCS Ist. G. Gaslini - Pediatric Hospital, interested in mass spectrometry applied to life sciences. Opinions are my own.

Fig. 4: Cross-omics network integration identifies coordinated EV proteomic-lipidomic modules.
(a) Schematic overview of the cross-omics integration strategy. Temporally regulated EV proteins (n = 875) and lipids (n = 100), identified by longitudinal linear mixed-effects modeling (LMM), were integrated using a protein-lipid correlation network based on Spearman correlations across paired longitudinal samples. Louvain community detection was applied to identify cross-omics modules, which were subsequently tested for associations with clinical outcomes. (b) Distribution of Spearman correlation coefficients for protein-lipid pairs retained in the network after multiple-testing correction and effect-size filtering (|ρ|>0.35). Positive and negative correlations are shown separately, highlighting a structured and bimodal distribution of inter-omic associations. (c) Protein and lipid composition of the five cross-omics modules (M1-M5). Bars indicate the relative contribution of proteins and lipids within each module, with total node counts reported above each bar. (d) Functional annotation of protein components within each module based on Gene Ontology Biological Process enrichment analysis. Bars represent −log10(FDR) values for selected enriched terms, highlighting distinct biological themes across modules, including protein turnover, immune-related processes, translation, and extracellular matrix remodeling.

Fig. 1: Longitudinal study design and plasma EV multi-omics profiling in very preterm infants. (a) Overview of the longitudinal cohort, sampling timeline, and analytical workflow. Plasma samples were collected from very preterm infants born at <32 weeks of gestational age (n = 16) at five time points from birth to term-equivalent age: at birth (T0), 48-72 hours of life (T1), 7 days of life (T2), 33 weeks postmenstrual age (PMA, T3), and 40 weeks PMA (T4, n = 10). EVs were enriched from plasma, followed by parallel isolation of EV-associated proteins and lipids from the same sample. Proteomic and lipidomic analyses were performed by LC-MS/MS, with raw data processed using dedicated pipelines prior to downstream statistical analyses. (b) Dynamic range of the plasma EV proteome, shown as a rank-abundance plot of median log₂ protein abundance across all samples, with selected EV-associated proteins highlighted. (c) Subcellular localization enrichment of the EV proteome based on the Jensen COMPARTMENTS database. Terms are ordered by the number of associated proteins along the x axis (protein count). Dot size reflects statistical significance (−log10 FDR), while color intensity indicates odds ratio. (d) Distribution of identified lipid species across major lipid classes, grouped into glycerophospholipids, glycerolipids, sphingolipids, fatty acyls, and sterol lipids.