Episode 13: Need information on transformation products? There's an app for that! - with Emma Palm

Live from SETAC Europe 2023 - Raoul Wolf hosts a series of ZeroPM podcasts. In this episode he interviews Emma Palm (Univesity of Luxembourg) about an app she developed to curate transformation product results, shiny TPs! An important consideration for understanding the hazards posed from different compounds is the transformation products (TPs) they form once they enter the environment. Unfortunately, there is much less information available for TPs than their parent compounds, while most of the information is text based in journal articles and databases. This makes it difficult to access and include into identification workflows based on high resolution mass spectrometry (HR-MS) and other analyses. One such database with text-based TP information is PubChem, which contains a lot of TP information, especially in the “Metabolism and Metabolites” section. In addition, PubChem runs a text mining software (LeadMine), which identifies chemical names in the text-based sections of the database. To better utilize this information, we have built a shiny app which will allow users to go through the text mined information in the Hazardous substance databank (HSDB) dataset to identify which of the extracted chemical names are TPs. After downloading the entire HSDB dataset and filtering based on the precursor compounds of interest, the app will display the structure of the precursor and the potential TP extracted via text mining, together with the HSDB text snippets to allow the user to identify whether the extracted structure is indeed a TP of the precursor. The information can then be saved in a table and downloaded for further analysis, or central upload as formal transformation reactions. In addition, the user has the option of saving the chemical names of TPs not identified by the text mining software together with their SMILES, to fill database gaps. The app has so far been applied to compounds from several datasets (28 from the NORMAN-SLE and 8 from the PubChem PFAS Tree, as well as 33 PBT compounds on the SIN-list). Of the 366 chemicals identified by the text mining software, 60 were new reactions that were not already available in the transformation products dataset in PubChem, the majority of which were found among the NORMAN and PBT compounds. Among these reactions were transformation products of azo dyes such as benzidine which has carcinogenic properties. In addition, some persistent substructures were identified in the TPs. This shows promise for expanding the current TP datasets available in databases such as PubChem as well as for improving TP suspect lists and furthering the understanding of the fate of environmental contaminants. See Emma's presentation at SETAC here: https://zenodo.org/record/7885830 To follow more of the work of ZeroPM, follow our ⁠Zenodo Community⁠ , youtube channel and our webpage to find all our resources  ⁠ZeroPM.eu⁠.