MSACL

We will be talking with Thomas Durant and Edward Lee about recent advances in machine learning (ML) for mass spectrometry (MS) data analysis. The primary focus will be the alignment of these two fields (ML and MS) and how this offers a promising synergy that can be used to optimize workflows, improve result quality, and enhance our understanding of high-dimensional datasets, as well as their inherent relationship with disease biology. We will also dig deeper to understand a basic overview of ML and an ML-based experiment. Overall, we will have a opportunity go through the fundamental principles of supervised ML, outline the steps that are classically involved in an ML-based experiment, and discuss the purpose of good ML practice in the context of a binary MS classification problem.

X linked disorders are considerably rare and research analysing human samples is under-represented. Research undertaken in this study used neural progenitor cells from an afflicted patient to begin exploring this rare disorder. While most experimental work focuses on the neurodevelopmental impacts of X-chromosome associated diseases, this work demonstrates that they should also be considered metabolic disorders owing to their perturbations on metabolite and lipid biochemistry. This work aims to use mass spectrometry to improve our understanding of these conditions and guide novel interventions by characterizing disease associated metabolic alterations.

Untargeted small molecule machine learning for MSI

Cameron Shedlock (until 14:00)

Mass spectrometry imaging returns large dimensionally complex datasets. Traditional analysis in imaging utilizes a mindset focused on target molecule identification that parses data with a narrow view of molecular exploration. These data analysis methods are specifically focused on exploring differences between a traditional organic acid matrix when compared to using nanoparticles for MSI, which results in very different ionization of small molecules. Analysis methods have taken a shotgun approach using both supervised and unsupervised machine learning to reveal critical trends in MSI datasets. A workflow is being prepared which enables select regions of interest to be compared using powerful machine learning algorithms to offer a holistic approach to data analysis and class comparison.

Mass spectrometry imaging using gold nanoparticles

Kate Stumpo, PhD

Mass spectrometry imaging (MSI) is a powerful analytical method for the simultaneous analysis of hundreds of compounds within a biological sample. Despite the broad applicability of this technique, there is a critical need for advancements in methods for small molecule detection. Some molecular classes of small molecules are more difficult than others to ionize, e.g., neurotransmitters (NTs). The chemical structure of NTs (i.e., primary, secondary, and tertiary amines) affects ionization and has been a noted difficulty in the literature. In order to achieve detection of NTs using MSI, strategies must focus on either changing the chemistry of target molecules to aid in detection or focus on new methods of ionization. This presentation will introduce a new method of ionization, using gold nanoparticles (AuNPs), and the bigger picture of NPs for MSI.

Lipids play a vital role in maintaining cellular functions. Altered lipid metabolism is currently considered a hallmark of many diseases, which highlights the importance of the characterization of lipid composition in understanding, diagnosing, and treating pathologies. Discrimination of isomeric species is challenging in lipidomics. In this talk, I will introduce the microdroplet electrochemical methods capable of resolving different types of isomers commonly encountered in lipid samples using electrospray ionization mass spectrometry. The methods take advantage of the voltage-controlled and dramatically accelerated electrochemical derivatization of lipid isomers in microdroplets to achieve structural elucidation. Applications of the electrochemical mass spectrometry methods in real sample analysis will also be included.

Surveying the food-associated compounds detected in clinical samples, we can differentiate patients with specific diet types, such as predominantly animal protein- versus plant-based diet. In addition, we can identify specific foods associated with clinical outcomes in disease cohorts. With its broad applications, we envision this approach becoming invaluable in nutrition research as well as many other fields once additional reference datasets become available.

This presentation will described the concept, methodology, validation and performance characteristics of a 10-µL, dilute-and-shoot, intact-protein mass spectrometric assay of albumin proteoforms called “Delta-S-Cys-Albumin” that serves as an endogenous marker of P/S exposure to thawed conditions based on the inexorable ex vivo S-cysteinylation (oxidizability) of albumin. The multi-reaction chemical mechanism that drives changes in albumin S-cysteinylation is known and the rate law for it was established and accurately modeled in P/S—enabling back-calculation of the time at which unknown P/S specimens have been exposed to the equivalent of room temperature. Results from blind challenges and two unanticipated case studies that revealed unexpected integrity problems in sets of nominally pristine P/S samples will be presented.

This session will outline the studies conducted and planned in order to complete validation of a mass spectrometry method in a CAP-accredited, clinical laboratory. While accuracy and precision and reportable range will be addressed, particular attention will be paid to the additional validation studies that should be conducted (matrix effect, injector stability, interferences, etc).

Unlike conventional immunoassays which provide simple (and often incorrect!) results, mass spectrometers provide a plethora of data for every peak- qualifier ions, internal standard areas, retention times, peak shapes, etc. This data, often termed metadata, can help clinical laboratories evaluate their peaks to ensure that they are reporting accurate results. Unfortunately, the process of evaluating this metadata can be a doubting challenge. This session will discover our efforts to move ahead with implementing manageable processes for evaluating the metadata necessary to ensure accurate results.

At the conclusion of this talk, participants should be able to…

1. Ask questions related to their own efforts to implement monitoring/assessing metadata

2. List the metadata that must be evaluated to ensure accurate results

3. Prepare quality assurance metrics for their laboratories utilizing metadata

What do we want? Good chromatography! How do we get it? We don’t know! Columns, maybe?

Good chromatography is essential to any LC-MS/MS method. Setting it up is often one of the earlier steps in developing a new method. While it can be an iterative process, setting up good chromatographic conditions at the start can set you up for success later on. This podcast will discuss our laboratories efforts to setup chromatographic conditions for the analysis of ethyl-glucuronide and buprenorphine (plus metabolites). We will present what we tried, why it seemed like it a good/bad idea, and along the way talk through some basic chromatography concepts. The goal is for attendees to be more informed when they approach setting up their own chromatographic methods.

This session is the second in a 4 part series in which Dr. Hayden will invite attendees to witness in real time his journey bringing mass spec testing to a clinical lab. During these interactive sessions, attendees will be encouraged to help troubleshoot, and offer advice as desired.

You can watch the recording of the first session: "Getting going with mass spectrometry: Josh installs a mass spectrometer".

Subsequent sessions anticipated include:

Part 3. Getting going with mass spectrometry: Josh tries to do sample preparation

Part 4. Getting going with mass spectrometry: Josh analyzes peaks

Prof. Hewison’s group is at the forefront of research linking vitamin D and the immune system, with implications for a wide range of clinical disorders including infectious, inflammatory and autoimmune disease. Current studies are focused on the analysis of vitamin D-insensitivity in T lymphocytes from the inflamed joints of patients with rheumatoid arthritis. This may provide an explanation for the limited success of vitamin D supplementation in some clinical trials. Other projects have explored the role of cell metabolism pathways in mediating the immunomodulatory effects of vitamin D, and the opportunities this may provide for improved therapeutic use of vitamin D. The Hewison group has also pioneered a range of studies to explore alternative markers of vitamin D ‘status’. This includes development of novel high throughput liquid chromatography-tandem mass spectrometry technology to measure multiple metabolites of vitamin D – the vitamin D metabolome – and analysis of the role of the serum vitamin D binding protein as a determinant of vitamin D bioavailability within the immune system.