Louise Malm PhD Student
Contact
Name and title: Louise MalmPhD Student
Workplace: Department of Chemistry Länk till annan webbplats.
Visiting address Room A315Svante Arrhenius väg 16 C
Postal address Kemikum10691 Stockholm
About me
PhD student in Anneli Kruve's group.
Obtained my MSc degree in Analytical Chemistry from Stockholm University in June 2022. The MSc thesis, was written in collaboration with University of Athens, NORMAN network and Quantem Analytics to conduct an interlaboratory comparison of semi-quantitative LC/ESI/HRMS analysis.
The project aim was to compare five semi-quantification methods by analysing spiked water samples. The thesis work included selecting the chemicals to spike water samples with, preparing and sending samples to 45 participating laboratories around the world, investigate the stability of selected chemicals, analyse the samples, set up a data collection platform, and analyse the collected results.
As a PhD student in the group of Anneli Kruve, I will be working on a project with the tentative title “Graph-based machine learning models for ionisation efficiency predictions in liquid chromatography electrospray ionisation high resolution mass spectrometry data to quantify emerging environmental contaminants”. In my research, I aim to develop high accuracy tools to quantify emerging contaminants without their analytical standards.
Will be teaching as laboratory assistant in the MSc level course “Advanced Separation Methods”.
The seemingly endless pollution of our environment can hardly have escaped anyone’s notice. Detection of chemicals with known or possible health effects are constantly being reported. The method of choice for analyzing environmental samples is liquid chromatography electrospray ionization high resolution mass spectrometry (LC/ESI/HRMS). The high resolving power of HRMS makes it possible to distinguish between very small masses. Together with the use of non-targeted screening, that is to analyze the sample without looking for any predetermined target chemicals, the identification of a wide range of contaminants has been made possible. However, to fully assess the risk any chemical poses, it is essential to know at which concentration it is found in the environment. To obtain this quantitative information, the analysis of analytical standards at several concentrations, at the same time as the analysis of the sample is needed to establish a calibration curve, from where the concentration can be determined. On the other hand, including analytical standards in the analysis, when you do not know what you are looking for, is quite the impossible task. As an alternative, the concentrations can be estimated; instead of using the calibration curves of each exact chemical that is tentatively identified, a calibration curve of another, so called calibration or surrogate standard, can be used. In this large interlaboratory study, we have compared five approaches to quantify chemicals in water samples. Three of the approaches use the calibration curve of a structurally similar or a close eluting surrogate standard, while two approaches use machine learning models for concentration estimations. The results clearly shows that the machine learning models provide more accurate concentration estimates, and that continued research is needed to further refine these methods.

