the Coronavirus Infectious Disease Ontology (CIDO) is a community-based ontology that supports coronavirus disease knowledge and data standardization, integration, sharing, and analysis. O ntologies, as the term is used in informatics, are structured vocabularies comprised of human-and computer-interpretable terms and relations that represent entities and relationships. Within informatics fields, ontologies play an important role in knowledge and data standardization, representation, integration, sharing and analysis. They have also become a foundation of artificial intelligence (AI) research. In what follows, we outline the Coronavirus Infectious Disease Ontology (CIDO), which covers multiple areas in the domain of coronavirus diseases, including etiology, transmission, epidemiology, pathogenesis, diagnosis, prevention, and treatment. We emphasize CIDO development relevant to COVID-19. Human coronaviruses have given rise to a series of major crises in global public health. Severe acute respiratory syndrome (SARS) emerged in China in November 2002, lasted for eight months and resulted in 8,098 confirmed human cases in 29 countries with 774 deaths (case-fatality rate: 9.6%) 1. Approximately ten years later in June 2012, the Middle East Respiratory Syndrome (MERS), another highly pathogenic coronavirus disease, was identified in Saudi Arabia. The MERS outbreak has caused 2,260 cases in 27 countries and 803 deaths (35.5%) 2. More recently, the World Health Organization (WHO) declared the Coronavirus Disease 2019 (COVID-19) outbreak as a pandemic on March 11, 2020, when there were 118,326 confirmed cases and 4,292 deaths. As of May 13, there have been over 4.4 million confirmed cases and over 295,000 deaths globally. Unfortunately, we still do not have available effective drugs and vaccines against these highly pathological coronaviruses. Extensive studies have been conducted on coronaviruses, the results of many of which exist in publicly available data repositories such as GEO 3. Publications concerning COVID-19 have exploded in recent months, and new clinical trials have been and are being conducted to develop drugs and vaccines against COVID-19, 1,430 of which have been registered in ClinicalTrials.

This Perspective provides examples of current and future applications of deep learning in pharmacogenomics, including: identification of novel regulatory variants located in noncoding domains of the genome and their function as applied to pharmacoepigenomics; patient stratification from medical records; and the mechanistic prediction of drug response, targets and their interactions. Deep learning encapsulates a family of machine learning algorithms that has transformed many important subfields of artificial intelligence over the last decade, and has demonstrated breakthrough performance improvements on a wide range of tasks in biomedicine. We anticipate that in the future, deep learning will be widely used to predict personalized drug response and optimize medication selection and dosing, using knowledge extracted from large and complex molecular, epidemiological, clinical and demographic datasets.

The actin cytoskeleton of rabbit proximal tubules was assessed by deoxyribonuclease (DNase) binding, sedimentability of detergent-insoluble actin, laser-scanning confocal microscopy, and ultrastructure during exposure to hypoxia, antimycin, or antimycin plus ionomycin. One-third of total actin was DNase reactive in control cells prior to deliberate depolymerization, and a similar proportion was unsedimentable from detergent lysates during 2.5 h at 100,000 g. Tubules injured by hypoxia or antimycin alone, without glycine, showed Ca(2+)-dependent pathology of the cytoskeleton, consisting of increases in DNase-reactive actin, redistribution of pelletable actin, and loss of microvilli concurrent with lethal membrane damage. In contrast, tubules similarly depleted of ATP and incubated with glycine showed no significant changes of DNase-reactive actin or actin sedimentability for up to 60 min, but, nevertheless, developed substantial loss of basal membrane-associated actin within 15 min and disruption of actin cores and clubbing of microvilli at durations > 30 min. These structural changes that occurred in the presence of glycine were not prevented by limiting Ca2+ availability or pH 6.9. Very rapid and extensive cytoskeletal disruption followed antimycin-plus-ionomycin treatment. In this setting, glycine and pH 6.9 decreased lethal membrane damage but did not ameliorate pathology in the cytoskeleton or microvilli; limiting Ca2+ availability partially protected the cytoskeleton but did not prevent lethal membrane damage. The data suggest that both ATP depletion-dependent but Ca(2+)-independent, as well as Ca(2+)-mediated, processes can disrupt the actin cytoskeleton during acute proximal tubule cell injury; that both types of change occur, despite protection afforded by glycine and reduced pH against lethal membrane damage; and that Ca(2+)-independent processes primarily account for prelethal actin cytoskeletal alterations during simple ATP depletion of proximal tubule cells.

Quantitative analysis of morphological changes in a cell nucleus is important for the understanding of nuclear architecture and its relationship with pathological conditions such as cancer. However, dimensionality of imaging data, together with a great variability of nuclear shapes, presents challenges for 3D morphological analysis. Thus, there is a compelling need for robust 3D nuclear morphometric techniques to carry out population-wide analysis. We propose a new approach that combines modeling, analysis, and interpretation of morphometric characteristics of cell nuclei and nucleoli in 3D. We used robust surface reconstruction that allows accurate approximation of 3D object boundary. Then, we computed geometric morphological measures characterizing the form of cell nuclei and nucleoli. Using these features, we compared over 450 nuclei with about 1,000 nucleoli of epithelial and mesenchymal prostate cancer cells, as well as 1,000 nuclei with over 2,000 nucleoli from serum-starved and proliferating fibroblast cells. Classification of sets of 9 and 15 cells achieved accuracy of 95.4% and 98%, respectively, for prostate cancer cells, and 95% and 98% for fibroblast cells. To our knowledge, this is the first attempt to combine these methods for 3D nuclear shape modeling and morphometry into a highly parallel pipeline workflow for morphometric analysis of thousands of nuclei and nucleoli in 3D.