The Short-Term Scientific Mission (STSM) allowed me to visit the Lipid Mediator Unit, Queen Mary University of London (UK), led by Prof. Jesmond Dalli. The visit had two main objectives. The first was to deepen my knowledge on the specialized pro-resolving mediators (SPMs), and specifically on their assessment method. The second objective was to learn how to implement artificial intelligence-driven approaches to identify biomarkers and look at pathway interactions.
As part of my postdoctoral project, I selected the participants from an ongoing cohort study for a case-control subset and sent their plasma samples to Prof. Dalli’s lab to analyse the SPM concentrations. During the STSM, I had the opportunity to see the facility in which the measurements were conducted and first-hand analyse some of the primary mass-spectrometry data output from a few batches of our samples. This gave me the opportunity to understand the process and the complexity of SPM identification and quantification, which mainly derive from their very low concentrations in human plasma. Furthermore, we discussed the best statistical analysis approaches to analyse our data. We explored a few artificial intelligence-driven approaches, which however were not feasible because of the nature of the data collected and specifically the large proportion of SMPs (~85%) falling below the lower limit of identification. Nonetheless, we identified a few relevant approaches that could help us explore the data, including principal component analysis (PCA), network analyses and conditional logistic regression.
