Today’s media places a large on social media reactions to news events. However, there is little
study on how accurate these reactions are to the reality, rather than to the same media stories about it. This study aims to try to see how accurately Tweets about COVID-19 reflect the reality of the situation. Since COVID-19 is a global crisis that spreads across geographic regions at different rates, it is a unique opportunity to study this type of behavior. Social media responses (such as tweets) can be classified as COVID-19 relevant or not. Correlating this against data on the spread of the disease, or against geographical increase in quarantine procedures, give insight into how social media responds to the effects of the virus.

Drug repurposing has the potential to expedite the discovery process by using commercially available compounds. Based on previous studies, we selected emodin, remdesivir, and hydroxychloroquine as candidates for our search. SARS-CoV-2 infects cells by using Angiotensin-Converting Enzyme 2 (ACE-2) as an entry receptor; emodin and hydroxychloroquine were suggested to be inhibitors of this interaction. Remdesivir acts as a ribonucleoside analog and interferes with the viral RNA polymerase of SARS-CoV-2. So far, we have discovered multiple ligands for each compound that had a higher binding affinity than their original candidates.

This project is jointly handled by computer science and biochemistry research students. We are working on a machine learning platform for rapid screening of libraries of chemical entities that have the potential of serving as inhibitors of the interaction between the coronavirus spike glycoprotein and the human Angiotensin-Converting Enzyme 2 (ACE2), which is the putative means of entry of SARS-CoV-2 into cells. Moreover, we are currently working towards building in methods for qualifying libraries of chemical entities in highly simplified molecular descriptors, which we envision will be greatly enabling in the efficiency by which compounds can be screened. This will eventually inform the future development of small molecule therapies. 

Electron-deficient Michael acceptors have been previously reported in application to irreversible, covalent inhibition of cysteine proteases. Here, we report a highly charge-dependent structure-activity relationship (SAR) of dehydroalanine-peptoid-based covalent inhibitors of the SARS-CoV-2 main protease, which is responsible for production of viral components, and we are close to completion of the SAR generated from an initial screen of 384 compounds, all of which can be chemically synthesized in < 10 steps. The initial SAR reported by our June preprint will serve as the basis for structure-informed and reactivity-guided design of next-generation protease inhibitors.  

Here, we use public data sources and computer modeling to build models that give projections for COVID-19 infection rates in certain geographic locations.  We will extend to other geographies or dig deeper to see what kind of policies at the county or city level seem to be affecting the decline the most. This work will be carried out with a combination of approaches in data science and computer science.  

ddRNAi, or DNA-directed RNA interference, is a gene-silencing method that uses DNA constructs to hijack the pre-existing animal cell’s RNA interference pathways. Here, we developed an RNA fragment that may silence a gene in the SARS-CoV-2 genome, which inhibits viral replication, thereby inhibiting the virus’ spread. We designed a novel ddRNAi strand that inhibits gene expression in SARS-CoV-2 infected cells. Our designed siRNA strand does not share any complete similarities with the known human genome and corresponds to the SARS-CoV-2 virus genome at 9841-9859 bp.