Bonicelli, A. | bioRxiv Science

bioRxiv2026

Integrated transcriptomic analysis identifies Paternally Expressed Gene 10 as a novel candidate biomarker associated with aggressive features in liposarcoma

Abstract Background Liposarcoma is a rare cancer, comprising four major histological subtypes: well-differentiated, dedifferentiated, myxoid and pleomorphic liposarcoma. Therapeutic options remain limited for advanced disease, and additional biomarkers that may support biological characterization, diagnosis, or risk stratification are needed. Methods Bulk RNA sequencing (RNA-seq) was performed on paired tumour and adjacent non-tumour tissues from an analysable subset of 16 patients (22 tumour specimens) within a clinicopathological cohort of 21 patients. Differential gene expression analysis, principal component analysis, gene set enrichment (GSEA) analysis and single-sample GSEA (ssGSEA) were used to define subtype-specific transcriptional programs and imprinted-gene dysregulation. PEG10 expression was validated by RT-qPCR and RNA fluorescence in situ hybridization (FISH), and its cellular localization was explored by combined FISH and immunofluorescence for P53, vimentin, Ki67, and CD34. PEG10 expression in liposarcoma was further investigated in an external single-cell RNA-sequencing dataset of dedifferentiated liposarcoma. An exploratory survival analysis was performed in the pan-sarcoma TCGA-SARC cohort. Results Transcriptomic profiling identified distinct subtype-specific gene expression patterns and broad dysregulation of imprinted genes in liposarcoma compared with adjacent non-tumour tissue. Among all the imprinted genes, PEG10 showed a marked tumour-to-normal expression difference and good discriminatory performance between tumour and normal tissues in the discovery cohort (AUC=0.867). In the TCGA sarcoma cohort high PEG10 expression was associated with shorter overall survival. Immuno-FISH and single-cell analyses revealed that PEG10 was predominantly detected in tumour and proliferating cell populations. Importantly, PEG10 expression was higher in high-grade tumours. In bulk RNA-seq data, PEG10 expression correlated with hypoxia, extracellular matrix remodelling and invasion ssGSEA signatures. In the external single-cell RNA-seq dataset PEG10high tumours showed a reduced immune cell infiltration. Conclusions This study provides an integrated transcriptomic characterisation of liposarcoma and identifies PEG10 as a candidate tumour-associated biomarker linked to aggressive disease features. The diagnostic and prognostic value of PEG10 requires validation in larger, independent, subtype-balanced liposarcoma cohorts, and its potential therapeutic relevance requires direct functional investigation.

genomics↗

bioRxiv2022

The "ForensOMICS" approach to forensic post-mortem interval estimation: combining metabolomics, lipidomics and proteomics for the analysis of human bone.

The combined use of multiple omics methods to answer complex system biology questions is growing in biological and medical sciences, as the importance of studying interrelated biological processes in their entirety is increasingly recognized. We applied a combination of metabolomics, lipidomics and proteomics to human bone to investigate the potential of this multi-omics approach to estimate the time elapsed since death (i.e., the post-mortem interval, PMI). This "ForensOMICS" approach has the potential to improve accuracy and precision of PMI estimation of skeletonized human remains, thereby helping forensic investigators to establish the timeline of events surrounding death. Anterior midshaft tibial bone was collected from four female body donors in a fresh stage of decomposition before placement of the bodies to decompose outdoors at the human taphonomy facility managed by the Forensic Anthropological Center at Texas State (FACTS). Bone samples were again collected at selected PMIs (219, 790, 834 and 872 days). Liquid chromatography mass spectrometry (LC-MS) was used to obtain untargeted metabolomic, lipidomic and proteomic profiles from the pre- and post-placement bone samples. Multivariate analysis was used to investigate the three omics blocks by means of Data Integration Analysis for Biomarker discovery using Latent variable approaches for Omics studies (DIABLO), to identify the reduced number of markers that could effectively describe post-mortem changes and classify the individuals based on their PMI. The resulting model showed that pre-placement bone metabolome, lipidome and proteome profiles were clearly distinguishable from post-placement profiles. Metabolites associated with the pre-placement samples, suggested an extinction of the energetic metabolism and a switch towards another source of fuelling (e.g., structural proteins). We were able to identify certain biomolecules from the three groups that show excellent potential for estimation of the PMI, predominantly the biomolecules from the metabolomics block. Our findings suggest that, by targeting a combination of compounds with different post-mortem stability, in future studies we could be able to estimate both short PMIs, by using metabolites and lipids, and longer PMIs, by including more stable proteins.

biochemistry↗