Table 1.
Overview of lipidomics and proteomics LC-MS analyses.
Table 2.
Untargeted lipid features identified via MS/MS fragmentation.
Fig 1.
Targeted lipidomics measurements per lipid class, with significantly enriched classes marked with red.
Targeted lipidomics data were grouped by lipid class and then evaluated for significance for high versus low risk MM (left) and RRMM versus NDMM (right) using enrichment analysis of fgsea R package. Lipid classes with adjusted P value < 0.05 are considered significantly different between the two groups (labelled red). LogFC, log fold change.
Table 3.
Reduced abundance of phosphatidylcholines (PC) in high risk and RRMM, measured by targeted lipidomics.
Fig 2.
Overlap between proteomics and transcriptomics data at the gene and pathway levels.
Proteomic level changes in RRMM compared to NDMM were evaluated against independent transcriptome data from the Multiple Myeloma Research Consortium reference collection. The graph shows the number of genes/proteins (left) or pathways (right) that are significantly different in the proteomics data (red bar), which also was significantly different in the transcriptome data (green bar), in the same direction (blue bar).
Fig 3.
Receiver operating characteristic (ROC) curve for correlated path classification model of lipid metabolic pathways based on transcriptome data for RRMM.
Diagnostic plot of the result from the path classification model for RRMM transcriptome data. ROC curves are shown for each component (M1, M2), which represent a path structure pattern. This gives information about which components is associated with RRMM and NDMM. A ROC curve with an AUC < 0.5 relates to RRMM. Conversely, ROC curve with AUC > 0.5 relates NDMM. Complete ROC represents the performance of the classifier using both components.
Fig 4.
Extracted correlated lipid metabolism path network for RRMM and NDMM patients.
A sub-network comprised of top 50 correlated paths based on gene expression in RRMM and NDMM was extracted from the lipid metabolism path network. Red and blue edges indicate exclusive correlation in RRMM and NDMM patients, respectively. Grey edges indicate correlation in both conditions.
Fig 5.
Proteomics results shown in the context of extracted lipid metabolism path network for RRMM and NDMM patients.
Proteomic data were projected on to the same network shown in Fig 4. Red nodes indicate up-regulation at protein level in RRMM compared to NDMM. Conversely, blue nodes indicated down-regulated proteins. Inset: PC metabolism pathways, showing expression correlation and proteomics up-regulation suggest active PC degradation in RRMM.