Recent advances in omics technologies have led to unprecedented efforts characterizing the molecular changes that underlie the development and progression of a wide array of complex human diseases, including cancer. As a result, multi-omics analyses—which take advantage of these technologies in genomics, transcriptomics, epigenomics, proteomics, metabolomics, and other omics areas—have been proposed and heralded as the key to advancing precision medicine in the clinic. In the field of precision oncology, genomics approaches, and, more recently, other omics analyses have helped reveal several key mechanisms in cancer development, treatment resistance, and recurrence risk, and several of these findings have been implemented in clinical oncology to help guide treatment decisions. However, truly integrated multi-omics analyses have not been applied widely, preventing further advances in precision medicine. Additional efforts are needed to develop the analytical infrastructure necessary to generate, analyze, and annotate multi-omics data effectively to inform precision medicine-based decision-making.

Metabolic diseases such as obesity and diabetes are associated with changes in high-density lipoprotein (HDL) particles, including changes in particle size and protein composition, often resulting in abnormal function. Recent studies suggested that patients with non-alcoholic fatty liver disease (NAFLD), including individuals with non-alcoholic steatohepatitis (NASH), have smaller HDL particles when compared to individuals without liver pathologies. However, no studies have investigated potential changes in HDL particle protein composition in patients with NAFLD, in addition to changes related to obesity, to explore putative functional changes of HDL which may increase the risk of cardiovascular complications. From a cohort of morbidly obese females who were diagnosed with simple steatosis (SS), NASH, or normal liver histology, we selected five matched individuals from each condition for a preliminary pilot HDL proteome analysis. HDL particles were enriched using size-exclusion chromatography, and the proteome of the resulting fraction was analyzed by liquid chromatography tandem mass spectrometry. Differences in the proteomes between the three conditions (normal, SS, NASH) were assessed using label-free quantitative analysis. Gene ontology term analysis was performed to assess the potential impact of proteomic changes on specific functions of HDL particles. Of the 95 proteins identified, 12 proteins showed nominally significant differences between the three conditions. Gene ontology term analysis revealed that severity of the liver pathology may significantly impact the anti-thrombotic functions of HDL particles, as suggested by changes in the abundance of HDL-associated proteins such as antithrombin III and plasminogen. The pilot data from this study suggest that changes in the HDL proteome may impact the functionality of HDL particles in NAFLD and NASH patients. These proteome changes may alter cardio-protective properties of HDL, potentially contributing to the increased cardiovascular disease risk in affected individuals. Further validation of these protein changes by orthogonal approaches is key to confirming the role of alterations in the HDL proteome in NAFLD and NASH. This will help elucidate the mechanistic effects of the altered HDL proteome on cardioprotective properties of HDL particles.

Shotgun proteomics utilizes a database search strategy to compare detected mass spectra to a library of theoretical spectra derived from reference genome information. As such, the robustness of proteomics results is contingent upon the completeness and accuracy of the gene annotation in the reference genome. For animal models of disease where genomic annotation is incomplete, such as non-human primates, proteogenomic methods can improve the detection of proteins by incorporating transcriptional data from RNA-Seq to improve proteomics search databases used for peptide spectral matching. Customized search databases derived from RNA-Seq data are capable of identifying unannotated genetic and splice variants while simultaneously reducing the number of comparisons to only those transcripts actively expressed in the tissue. We collected RNA-Seq and proteomic data from 10 vervet monkey liver samples and used the RNA-Seq data to curate sample-specific search databases which were analyzed in the program Morpheus. We compared these results against those from a search database generated from the reference vervet genome. A total of 284 previously unannotated splice junctions were predicted by the RNA-Seq data, 92 of which were confirmed by peptide spectral matches. More than half (53/92) of these unannotated splice variants had orthologs in other non-human primates, suggesting that failure to match these peptides in the reference analyses likely arose from incomplete gene model information. The sample-specific databases also identified 101 unique peptides containing single amino acid substitutions which were missed by the reference database. Because the sample-specific searches were restricted to actively expressed transcripts, the search databases were smaller, more computationally efficient, and identified more peptides at the empirically derived 1 % false discovery rate. Proteogenomic approaches are ideally suited to facilitate the discovery and annotation of proteins in less widely studies animal models such as non-human primates. We expect that these approaches will help to improve existing genome annotations of non-human primate species such as vervet.

Background: Reliable and effective label-free quantification (LFQ) analyses are dependent not only on the method of data acquisition in the mass spectrometer, but also on the downstream data processing, including software tools, query database, data normalization and imputation. In non-human primates (NHP), LFQ is challenging because the query databases for NHP are limited since the genomes of these species are not comprehensively annotated. This invariably results in limited discovery of proteins and associated Post Translational Modifications (PTMs) and a higher fraction of missing data points. While identification of fewer proteins and PTMs due to database limitations can negatively impact uncovering important and meaningful biological information, missing data also limits downstream analyses (e.g., multivariate analyses), decreases statistical power, biases statistical inference, and makes biological interpretation of the data more challenging. In this study we attempted to address both issues: first, we used the MetaMorphues proteomics search engine to counter the limits of NHP query databases and maximize the discovery of proteins and associated PTMs, and second, we evaluated different imputation methods for accurate data inference. We used a generic approach for missing data imputation analysis without distinguising the potential source of missing data (either non-assigned m/z or missing values across runs). Results: Using the MetaMorpheus proteomics search engine we obtained quantitative data for 1622 proteins and 10,634 peptides including 58 different PTMs (biological, metal and artifacts) across a diverse age range of NHP brain frontal cortex. However, among the 1622 proteins identified, only 293 proteins were quantified across all samples with no missing values, emphasizing the importance of implementing an accurate and statiscaly valid imputation method to fill in missing data. In our imputation analysis we demonstrate that Single Imputation methods that borrow information from correlated proteins such as Generalized Ridge Regression (GRR), Random Forest (RF), local least squares (LLS), and a Bayesian Principal Component Analysis methods (BPCA), are able to estimate missing protein abundance values with great accuracy. Conclusions: Overall, this study offers a detailed comparative analysis of LFQ data generated in NHP and proposes strategies for improved LFQ in NHP proteomics data.

Maternal obesity (MO) during pregnancy is linked to increased and premature risk of age‐related metabolic diseases in the offspring. However, the underlying molecular mechanisms still remain not fully understood. Using a well‐established nonhuman primate model of MO, we analyzed tissue biopsies and plasma samples obtained from post‐pubertal offspring (3–6.5 y) of MO mothers ( n = 19) and from control animals born to mothers fed a standard diet (CON, n = 13). All offspring ate a healthy chow diet after weaning. Using untargeted gas chromatography‐mass spectrometry metabolomics analysis, we quantified a total of 351 liver, 316 skeletal muscle, and 423 plasma metabolites. We identified 58 metabolites significantly altered in the liver and 46 in the skeletal muscle of MO offspring, with 8 metabolites shared between both tissues. Several metabolites were changed in opposite directions in males and females in both liver and skeletal muscle. Several tissue‐specific and 4 shared metabolic pathways were identified from these dysregulated metabolites. Interestingly, none of the tissue‐specific metabolic changes were reflected in plasma. Overall, our study describes characteristic metabolic perturbations in the liver and skeletal muscle in MO offspring, indicating that metabolic programming in utero persists postnatally, and revealing potential novel mechanisms that may contribute to age‐related metabolic diseases later in life.

Gas chromatography–coupled mass spectrometry (GC–MS) has been used in biomedical research to analyze volatile, non-polar, and polar metabolites in a wide array of sample types. Despite advances in technology, missing values are still common in metabolomics datasets and must be properly handled. We evaluated the performance of ten commonly used missing value imputation methods with metabolites analyzed on an HR GC–MS instrument. By introducing missing values into the complete (i.e., data without any missing values) National Institute of Standards and Technology (NIST) plasma dataset, we demonstrate that random forest (RF), glmnet ridge regression (GRR), and Bayesian principal component analysis (BPCA) shared the lowest root mean squared error (RMSE) in technical replicate data. Further examination of these three methods in data from baboon plasma and liver samples demonstrated they all maintained high accuracy. Overall, our analysis suggests that any of the three imputation methods can be applied effectively to untargeted metabolomics datasets with high accuracy. However, it is important to note that imputation will alter the correlation structure of the dataset and bias downstream regression coefficients and p-values.