Journal of the American Society for Mass Spectrometry

Evaluation of Serum Proteome Sample Preparation Methods to Support Clinical Proteomics Applications

Serum contains several proteins that are associated with disease-related processes. Mass spectrometry (MS)-based proteomics approaches greatly facilitate serum protein biomarker development. However, the serum proteome complexity presents a technical challenge for the accurate, sensitive, and reproducible quantification of proteins by MS. Thus, efficient sample preparation methods are of critical importance for serum proteome analyses. In this study, we evaluated the technical performance of two serum proteome sample preparation methods using sera from patients with high-grade serous ovarian cancer and patients with benign nongynecological conditions with a goal of providing insight into their compatibility with clinical proteomics workflows. One method entailed the use of immobilized trypsin (SMART Digest Trypsin) with RapiGest SF, an acid-labile surfactant designed to enhance the in-solution enzymatic digestion of proteins. The other method incorporated a commercially available sample preparation kit, iST-BCT, which contains standardized reagents. Significantly higher protein sequence coverage, albeit with lower digestion efficiency, was obtained with the immobilized trypsin + RapiGest SF workflow, whereas the iST-BCT workflow was quicker and had marginally better reproducibility. Protein relative abundance analysis revealed that the serum proteomes clustered primarily by the sample processing workflow and secondarily by disease state. We conducted a time course study to determine whether differences in the relative abundance of diagnostic high-grade serous ovarian cancer serum protein biomarker candidates were biased according to the duration of enzymatic digestion. Our results highlight the importance of optimizing enzymatic digestion kinetics according to the peptide targets of interest while considering the sensitivity of the downstream analytical method utilized in clinical proteomics workflows designed to measure biomarkers.

Automated Machine Learning and Explainable AI (AutoML-XAI) for Metabolomics: Improving Cancer Diagnostics

Metabolomics generates complex data necessitating advanced computational methods for generating biological insight. While machine learning (ML) is promising, the challenges of selecting the best algorithms and tuning hyperparameters, particularly for nonexperts, remain. Automated machine learning (AutoML) can streamline this process; however, the issue of interpretability could persist. This research introduces a unified pipeline that combines AutoML with explainable AI (XAI) techniques to optimize metabolomics analysis. We tested our approach on two data sets: renal cell carcinoma (RCC) urine metabolomics and ovarian cancer (OC) serum metabolomics. AutoML, using Auto-sklearn, surpassed standalone ML algorithms like SVM and k-Nearest Neighbors in differentiating between RCC and healthy controls, as well as OC patients and those with other gynecological cancers. The effectiveness of Auto-sklearn is highlighted by its AUC scores of 0.97 for RCC and 0.85 for OC, obtained from the unseen test sets. Importantly, on most of the metrics considered, Auto-sklearn demonstrated a better classification performance, leveraging a mix of algorithms and ensemble techniques. Shapley Additive Explanations (SHAP) provided a global ranking of feature importance, identifying dibutylamine and ganglioside GM(d34:1) as the top discriminative metabolites for RCC and OC, respectively. Waterfall plots offered local explanations by illustrating the influence of each metabolite on individual predictions. Dependence plots spotlighted metabolite interactions, such as the connection between hippuric acid and one of its derivatives in RCC, and between GM3(d34:1) and GM3(18:1_16:0) in OC, hinting at potential mechanistic relationships. Through decision plots, a detailed error analysis was conducted, contrasting feature importance for correctly versus incorrectly classified samples. In essence, our pipeline emphasizes the importance of harmonizing AutoML and XAI, facilitating both simplified ML application and improved interpretability in metabolomics data science.

Proteomic Analysis Reveals Low-Dose PARP Inhibitor-Induced Differential Protein Expression in BRCA1-Mutated High-Grade Serous Ovarian Cancer Cells

High-grade serous ovarian cancer (HGSOC) is the most common form of ovarian cancer diagnosed in patients worldwide. Patients with

Applying Multimodal Mass Spectrometry to Image Tumors Undergoing Ferroptosis Following In Vivo Treatment with a Ferroptosis Inducer

Epithelial ovarian cancer (EOC) is the most common form of ovarian cancer. The poor prognosis generally associated with this disease has led to the search for improved therapies such as ferroptosis-inducing agents. Ferroptosis is a form of regulated cell death that is dependent on iron and is characterized by lipid peroxidation. Precise mapping of lipids and iron within tumors exposed to ferroptosis-inducing agents may provide insight into processes of ferroptosis

Spatially Controlled Molecular Analysis of Biological Samples Using Nanodroplet Arrays and Direct Droplet Aspiration

Mass spectrometry (MS) has emerged as a valuable technology for molecular and spatial evaluation of biological samples. Ambient ionization MS techniques, in particular, allow direct analysis of tissue samples with minimal pretreatment. Here, we describe the design and optimization of an alternative ambient liquid extraction MS approach for metabolite and lipid profiling and imaging from biological samples. The system combines a piezoelectric picoliter dispenser to form solvent nanodroplets onto the sample surface with controlled and tunable spatial resolution and a conductive capillary to directly aspirate/ionize the nanodroplets for efficient analyte transmission and detection. Using this approach, we performed spatial profiling of mouse brain tissue sections with different droplet sizes (390, 420, and 500 μm). MS analysis of normal and cancerous human brain and ovarian tissues yielded rich metabolic profiles that were characteristic of disease state and enabled visualization of tissue regions with different histologic composition. This method was also used to analyze the lipid profiles of human ovarian cell lines. Overall, our results demonstrate the capabilities of this system for spatially controlled MS analysis of biological samples.