KNIME Data Talks: R&D in Life Sciences
Presentations from KNIME Data Talks on November 4, 2021.
Pushing the Boundaries of Data Integration with KNIME
Eric understands the importance of carefully blending, curating, and standardizing data from various sources to create insightful medical chemistry dashboards. In this talk, he explains how he used KNIME as a back-end to create 70 different dashboards used by more than 300 chemists. The flexibility of KNIME for collecting data from different sources enabled him to achieve this with limited resources. He used KNIME workflows to extract knowledge from heterogeneous models and aggregate large volumes of data (>450k compounds) before passing it over to TIBCO Spotfire for visualization.
Speaker: Eric Vangrevelinghe (Novartis Pharma AG)
Implementing a Cheminformatics Workflow Environment from Scratch in Medium-Sized Pharma
Dora shares her experiences with implementing a cheminformatics pipeline - using KNIME - from scratch. She describes the technical aspects of the implementation and deep dives into an example workflow. While explaining this workflow step-by-step, she shares some tips and tricks as well as some challenges she encountered and how they were overcome.
Speaker: Dora Šribar (Nuvisan ICB)
Using Machine Learning Techniques to Predict OH-Progesterone Perturbant Chemicals
Mario Negri Research institute shows how they use KNIME to generate models to predict steroidogenesis disruption and generate toxicological profiles of compounds. The models trained in KNIME are deployed to the VEGA platform, which collects various tools to evaluate chemical hazards. The steps of the workflow comprise data preparation and curation, calculation of chemical descriptors and subsequent feature reduction to obtain the best models. These are then made available on the VEGA Hub for users.
Speaker: Marco Marzo (Istituto di Ricerche Farmacologiche Mario Negri)