ABSTRACT
Due to the COVID-19 pandemic, the demand for distance learning has significantly increased in higher education institutions. This type of learning is usually supported by Web-based learning systems such as Massive Open Online Courses (Coursera, edX, etc.) and Learning Management Systems (Moodle, Blackboard-Learn, etc.). However, in this remote context, students often lack feedback and support from educational staff, especially when they face difficulties or challenges. For that reason, this work presents a Prediction-Intervention approach that (a) predicts students who present difficulties during an online learning course, based on two main learning indicators, namely engagement and performance rates, and (b) offers immediate support to students, tailored to the problem they are facing. To predict students' issues, our approach considers ten machine learning algorithms of different types (standalone, ensemble, and deep learning) which are compared to determine the best performing ones. It has been experimented with a dataset collected from the Blackboard-Learn platform utilized in an engineering school called ESIEE-IT in France during 2021-2022 academic year, showing thus quite promising results. © 2022 IEEE.
ABSTRACT
The Coronavirus disease 2019 (Covid-2019) pandemic currently provokes a global health and economic crisis due to the generalized spread of severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2). SARS-CoV-2 is a large, enveloped and positive sense single stranded RNA virus. The SARS-CoV-2 genome encodes 16 non-structural proteins (nsp 1-16), forming a large membrane bound replication complex. The largest protein of this complex is nsp3, a multi-domain protein that contain a well conserved Macro domain (also called X domain or ADP-ribose phosphatase domain). The Macro domain can bind to mono-ADP-ribose (MAR) and poly-ADP-ribose (PAR) in their free form or conjugated to a protein or RNA substrates. Macro domains also carry hydrolase activities including de-MARylation and de-PARylation implicated in the inflammation process and the regulation of innate immunity. Herein, we report a mutagenesis study focusing on SARS-CoV-2 F156 and SARSCoV N157 residues, stipulated important for ADP-ribose orientation within the binding clef. Our data suggest that the exchange of these residues or their substitution to alanine slightly influence ADP-ribose binding, but drastically impact Macro domain de-MARylation activity.