HARMLESS @ QSAR 2023
The 20th International Workshop on Quantitative Structure-Activity Relationships in Environmental and Health Sciences (QSAR2023) took place on 5-9 June 2023 in Copenhagen, Denmark. The workshop provided a framework for the exchange of information and views on emerging research results, tools and regulatory use related to (Q)SARs and data analysis for many types of chemical substances, including pharmaceuticals. It was the perfect forum for meeting experts in development and application of (Q)SARs, share results and experiences, and participate in lively discussions.
The event, organised by DTU (Technical University of Denmark), brought together 116 people from different stakeholder groups, from QSAR developers to users from academia, regulators, as well as industry from all over the world.
The scope of the workshop involved the following active research areas:
- Development and validation of (Q)SARs, various endpoints
- QSARs and Big data / bioinformatics
- Screening / prioritization and data gap filling including by AOP / IATA
- Other application of (Q)SARs / tools
- (Q)SAR and read-across and grouping
- (Q)SARs and toxicokinetic modelling
- Quantification of data and model uncertainties
- Descriptor and modelling methods developments
- Emerging technologies, challenges, QAAR, UVCBs / mixtures, nano, green chemistry / safe-by-design etc.
As toxicity assessment and prioritization was one of the focus topics of QSAR 2023 and is one of the goals of HARMLESS, the project had an active role in the event. Gergana Tancheva, from IDEA, presented a poster in the Session 5 (Screening and prioritization). In her poster, she explains an automated workflow that preprocesses data and calculates Tox5 scores from raw HTS data. A new Python module is developed for collecting and annotating raw data, normalizing the data, and calculating dose-response metrics. The module utilizes the ToxPi-R library and follows the original Tox5 approach developed by Karolinska Institute and Department of Biological Toxicology of Misvik. It can be used independently or as part of the Orange workflow, which is an open-source system for visual programming and machine learning. The Orange workflow includes various widgets for data normalization, dose-response calculation, Tox5 scoring, ranking, and visualization of toxicity scores for each material. The Python module and Orange workflow extend the eNanoMapper FAIRification workflow by facilitating the FAIRification (i.e., making data Findable, Accessible, Interoperable, and Reusable) of HTS data. The resulting FAIR data, including raw and interpreted data (scores), is provided in a machine-readable format, which can be distributed as a data archive or integrated into the eNanoMapper database. The poster is available here – DOI: 10.5281/zenodo.8029959
By sharing our work at the poster session, we disseminated our developments and they raised awareness and attracted attention among attendees, particularly in the area of FAIRification of HTS data, ultimately promoting the project’s goals, results and potential impact. The conference was an opportunity for knowledge sharing, collaboration and increased visibility in the QSAR research community. Additionally, the participation of regulators at the event, such as ECHA, EPA, FDA, BfR, interested in the safety data generated and the use of HTS biological data for efficient clustering, ranking and prioritization, was very relevant for the project. Moreover, we were able to check the actuality of HARMLESS’ developments in comparison with other industries in this field.