Coast e-lecture: "What can quantitative non-targeted LC/HRMS analysis reveal about samples and analysis methods?" (Lecture)
- Tuesday 12 May 2020Add to my calendar
- from 11:00
prof. Anneli Kruve (Stockholm University, Sweden)
Because of the Covid-19 pandemic and the measures taken by the government, the FAST conference was canceled as a physical meeting in May 2020.
Instead we announce:
* FAST KEYNOTE E-LECTURE: May 12th, 2020, 11:00 -12:00
The applicability of suspect and non-targeted screening with LC/ESI/MS in environmental monitoring is growing fast. The non-targeted analysis allows detecting hundreds or thousands of features simultaneously; however, the relevance of the features depends on the structure behind the features, on the potential effect (toxicity) of the species, and on the concentration of the species. Therefore, one of the most significant obstacles for non-targeted LC/HRMS screening has been the inability to provide quantitative information. To overcome this issue, we have developed a quantification approach based on in silico predicted electrospray ionization efficiencies.
The ionization efficiency values are predicted based on 2D descriptors with the random forest algorithm. The model was developed based on ~6000 ionization values and accounts for both analyte and eluent properties. The model can be transferred between different labs with the aid of a small set of calibration compounds. The method has been tested for quantification in a number of applications, including a general non-targeted screening of surface water as well as a screening of pesticides and their transformation products in surface water, groundwater and foodstuff. In all of the applications, an average concentration prediction accuracy of ~3× has been observed. This means that the reported concentration of 3 ppm is likely to be in the range of 1 to 9 ppm (3 ppm/3 times to 3 ppm x 3 times).
We have also used this method to quantitatively compare the results of suspect screening of different laboratories. Based on the suspect screening of pesticides in cereals we have seen that the concentrations estimated in two independent labs are in good agreement. An R2 of 0.85 was observed, along with a concentration disparity of 3.2× (on average). Based on the concentrations estimated by one of the labs we were able to identify the root cause of inconsistency in the tentatively identified compounds between the two labs: one of the labs used a method with significantly lower LoD values.
In case you would like to attend the free e-lecture please send an email to email@example.com with the subject “FAST e-Lecture May 12” before May 7th.
dr. Jeroen Jansen