Jaqpot - An open-source web platform for creating, using, testing and sharing predictive models in nano-informatics
Nano WG June 20, 2019
Computational models that predict physicochemical properties, biokinetics or adverse biological effects of Engineered Nanomaterials (ENMs) are becoming increasingly important to support risk assessment and safety-by-design. This is primarily due to cost-saving and reduction of attrition rates, since market candidates can be assessed early-on in the development process. ENMs predicted to be toxic or not possessing the required properties can be discarded before a significant amount of effort has been invested, and most significantly, before time-consuming and expensive experimental tests have been carried out.
Over the last decade, a growing number of specific to ENMs Quantitative-Structure-Activity-Relationship (QSAR) and biokinetics models have been presented in the literature. However, these models are fragmented in the scientific literature and in many cases the information provided in the research papers is not adequate to reproduce, test and use the models.
In this seminar, we will present Jaqpot, a web platform that has been developed by our group at NTUA, which allows to systematically produce, collect, organize, validate, store and share predictive models relevant to the nanosafety science. The goal is to create a central repository of predictive nanoinformatics models with ready-to-use graphical user interfaces (GUIs). We will briefly refer to the software technologies used for developing Jaqpot, i.e. the integration of many microservices using Docker containers and RESTful APIs. We will focus on Jaqpot from the end-user point of view, i.e. how a user can develop his/her own model, deploy it in the Jaqpot infrastructure, enter metadata and ontological information about the descriptors and the end-point(s) to be predicted, include QMRF documentations and PMML representations, validate the models and use them for calculating predictions, comment on the model performances and share the models across and within virtual organizations.
Cite this work
Researchers should cite this work as follows:
National Cancer Institute