Comparing machine learning screening approaches using clinical data and cytokine profiles for COVID-19 in resource-limited and resource-abundant settings.

Accurate screening of COVID-19 infection status for symptomatic patients is a critical public health task. Although molecular and antigen tests now exist for COVID-19, in resource-limited settings, screening tests are often not available. Furthermore, during the early stages of the pandemic tests were not available in any capacity. We utilized an automated machine learning (ML) approach to train and evaluate thousands of models on a clinical dataset consisting of commonly available clinical and laboratory data, along with cytokine profiles for patients (n = 150). These models were then further tested for generalizability on an out-of-sample secondary dataset (n = 120). We were able to develop a ML model for rapid and reliable screening of patients as COVID-19 positive or negative using three approaches: commonly available clinical and laboratory data, a cytokine profile, and a combination of the common data and cytokine profile. Of the tens of thousands of models automatically tested for the three approaches, all three approaches demonstrated > 92% sensitivity and > 88 specificity while our highest performing model achieved 95.6% sensitivity and 98.1% specificity. These models represent a potential effective deployable solution for COVID-19 status classification for symptomatic patients in resource-limited settings and provide proof-of-concept for rapid development of screening tools for novel emerging infectious diseases.

The ChatGPT conundrum: Human-generated scientific manuscripts misidentified as AI creations by AI text detection tool.

AI Chat Bots such as ChatGPT are revolutionizing our AI capabilities, especially in text generation, to help expedite many tasks, but they introduce new dilemmas. The detection of AI-generated text has become a subject of great debate considering the AI text detector's known and unexpected limitations. Thus far, much research in this area has focused on the detection of AI-generated text; however, the goal of this study was to evaluate the opposite scenario, an AI-text detection tool's ability to discriminate human-generated text. Thousands of abstracts from several of the most well-known scientific journals were used to test the predictive capabilities of these detection tools, assessing abstracts from 1980 to 2023. We found that the AI text detector erroneously identified up to 8% of the known real abstracts as AI-generated text. This further highlights the current limitations of such detection tools and argues for novel detectors or combined approaches that can address this shortcoming and minimize its unanticipated consequences as we navigate this new AI landscape.

Ynol Ethers as Ketene Equivalents in Rhodium-Catalyzed Intermolecular [5 + 2] Cycloaddition Reactions.

The previously unexplored metal-catalyzed [5 + 2] cycloadditions of vinylcyclopropanes (VCPs) and electron-rich alkynes (ynol ethers) have been found to provide a highly efficient, direct route to dioxygenated seven-membered rings, a common feature of numerous natural and non-natural targets and building blocks for synthesis. The reactions proceed in high yield at room temperature and tolerate a broad range of functionalities. Substituted VCPs were found to react with high regioselectivity.

Visible-light photocatalyzed cross-linking of diacetylene ligands by quantum dots to improve their aqueous colloidal stability.

Ligand cross-linking is known to improve the colloidal stability of nanoparticles, particularly in aqueous solutions. However, most cross-linking is performed chemically, in which it is difficult to limit interparticle cross-linking, unless performed at low concentrations. Photochemical cross-linking is a promising approach but usually requires ultraviolet (UV) light to initiate. Using such high-energy photons can be harmful to systems in which the ligand-nanoparticle bond is fairly weak, as is the case for the commonly used semiconductor quantum dots (QDs). Here, we introduce a novel approach to cross-link thiolated ligands on QDs by utilizing the photocatalytic activity of QDs upon absorbing visible light. We show that using visible light leads to better ligand cross-linking by avoiding the problem of ligand dissociation that occurs upon UV light exposure. Once cross-linked, the ligands significantly enhance the colloidal stability of those same QDs that facilitated cross-linking.

Optimization of peptidyl allyl sulfones as clan CA cysteine protease inhibitors.

This research investigates the synthesis and inhibitory potency of a series of novel dipeptidyl allyl sulfones as clan CA cysteine protease inhibitors. The structure of the inhibitors consists of a R(1)-Phe-R(2)-AS-Ph scaffold (AS = allyl sulfone). R(1) was varied with benzyloxycarbonyl, morpholinocarbonyl, or N-methylpiperazinocarbonyl substituents. R(2) was varied with either Phe of Hfe residues. Synthesis involved preparation of vinyl sulfone analogues followed by isomerization to allyl sulfones using n-butyl lithium and t-butyl hydroperoxide. Sterics, temperature and base strength were all factors that affected the formation and stereochemistry of the allyl sulfone moiety. The inhibitors were assayed with three clan CA cysteine proteases (cruzain, cathepsin B and calpain I) as well as one serine protease (trypsin). The most potent inhibitor, (E)-Mu-Phe-Hfe-AS-Ph, displayed at least 10-fold selectivity for cruzain over clan CA cysteine proteases cathepsin B and calpain I with a (kobs)/[I] of 6080 ± 1390 M(-1)s(-1).