An Analysis of Deep Learning Models to Diagnose COVID-19 Using Radiography Images

Coronavirus(COVID-19) has created havoc for humanity by causing millions of deaths, adverse effects on physical and mental health and disastrous economic destruction. To aid with fast and efficient detection of the virus which allows for timely treatment of the patients, we have conducted this research work. In this work we have experimented and analyzed all the pre-trained and well known models. The performance of these various models in the detection of COVID-19 from Chest X-Ray images and their comparative study is depicted in this work. The best performing models were InceptionV3 (99.78%), InceptionResNetV2 (99.56%), DenseNet121 (99.34%), DensetNet169 (99.24%), DenseNet201 (99.34%) and Xception (99.12%). © 2022 IEEE.

Emergence of Unique SARS-CoV-2 ORF10 Variants and Their Impact on Protein Structure and Function (preprint)

The devastating impact of the ongoing coronavirus disease 2019 (COVID-19) on public health, caused by the Severe Acute Respiratory Syndrome Coronavirus 2 (SARS-CoV-2) has made fighting of the COVID-19 pandemic is a top priority in medical research and pharmaceutical development. Surveillance of SARS-CoV-2 mutations is essential for the comprehension of SARS-CoV-2 variant diversity and their impact on virulence and pathogenicity. The SARS-CoV-2 open reading frame 10 (ORF10) protein interacts with multiple human proteins CUL2, ELOB, ELOC, MAP7D1, PPT1, RBX1, THTPA, TIMM8B, and ZYG11B expressed in the lung tissues. Mutations and co-mutations in the emerging SARS-CoV-2 ORF10 variants are expected to impact the severity of the virus and its associated consequences. In this article, We highlight 128 single mutations and 35 co-mutations in the unique SARS-CoV-2 ORF10 variants in this article. The possible predicted effects of these mutations and co-mutations on the secondary structure of ORF10 variants and host protein interactomes are presented. The findings highlight the possible effects of mutations and co-mutations on the emerging 140 ORF10 unique variants from secondary structure and intrinsic protein disorder perspectives.

Non-Uniform Aspects of SARS-CoV-2 Intraspecies Evolution Reopen Questions on Its Origin (preprint)

Several hypotheses have been presented on the origin of severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) from its identification as the agent causing the current coronavirus disease 19 (COVID-19) pandemic. So far, no hypothesis has managed to identify the origin, and the issue has resurfaced. Here we have unfolded a pattern of distribution of several mutations in the SARS-CoV-2 proteins across different continents comprising 24 geo-locations. The results showed an evenly uneven distribution of unique protein variants, distinct mutations, unique frequency of common conserved residues, and mutational residues across the 24 geo-locations. Furthermore, ample mutations were identified in the evolutionarily conserved invariant regions in the SARS-CoV-2 proteins across almost all geo-locations we have considered. This pattern of mutations potentially breaches the law of evolutionary conserved functional units of the beta-coronavirus genus. These mutations may lead to several novel SARS-CoV-2 variants with a high degree of transmissibility and virulence. A thorough investigation on the origin and characteristics of SARS-CoV-2 needs to be conducted in the interest of science and to be prepared to meet the challenges of potential future pandemics.

An Issue of Concern: Unique Truncated ORF8 Protein Variants of SARS-CoV-2 (preprint)

Open reading frame 8 (ORF8) protein is one of the most evolving accessory proteins in severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2), the causative agent of coronavirus disease 2019 (COVID-19). It was previously reported that the ORF8 protein inhibits presentation of viral antigens by the major histocompatibility complex class I (MHC-I) and interacts with host factors involved in pulmonary inflammation. The ORF8 protein assists SARS-CoV-2 to evade immunity and replication. Among many contributing mutations, Q27STOP, a mutation in the ORF8 protein defines the B.1.1.7 lineage of SARS-CoV-2, which is engendering the second wave of COVID-19. In the present study, 47 unique truncated ORF8 proteins (T-ORF8) due to the Q27STOP mutations were identified among 49055 available B.1.1.7 SARS-CoV-2 sequences. The results show that only one of the 47 T-ORF8 variants spread to over 57 geo-locations in North America, and other continents which includes Africa, Asia, Europe and South America. Based on various quantitative features such as amino acid homology, polar/non-polar sequence homology, Shannon entropy conservation, and other physicochemical properties of all specific 47 T-ORF8 protein variants, a collection of nine possible T-ORF8 unique variants were defined. The question of whether T-ORF8 variants work similarly to ORF8 has yet to be investigated. A positive response to the question could exacerbate future COVID-19 waves, necessitating severe containment measures.

Carbon-Based Nanomaterials: Promising Antiviral Agents to Combat COVID-19 in the Microbial Resistant Era (preprint)

Therapeutic options for the highly pathogenic human Severe Acute Respiratory Syndrome-Coronavirus 2 (SARS-CoV-2) causing the current pandemic Coronavirus disease (COVID-19) are urgently needed. COVID-19 is associated with viral pneumonia and acute respiratory distress syndrome causing significant morbidity and mortality. The proposed treatments for COVID-19, such as hydroxychloroquine, remdesivir and lopinavir/ritonavir, have shown little or no effect in the clinic. Additionally, bacterial and fungal pathogens contribute to the SARS-CoV-2 mediated pneumonia disease complex. The antibiotic resistance in pneumonia treatment is increasing at an alarming rate. Therefore, carbon-based nanomaterials (CBNs), such as fullerene, carbon dots, graphene, and their derivatives constitute a promising alternative due to their wide-spectrum antimicrobial activity, biocompatibility, biodegradability and capacity to induce tissue regeneration. Furthermore, the antimicrobial mode of action is mainly physical (e.g. membrane distortion), which is characterized by a low risk of antimicrobial resistance. In this review, we evaluated the literature on the antiviral activity and broad-spectrum antimicrobial properties of CBNs. CBNs had antiviral activity against 12 enveloped positive-sense single-stranded RNA viruses similar to SARS-CoV-2. CBNs with low or no toxicity to the humans are promising therapeutics against COVID-19 pneumonia complex with other viruses, bacteria and fungi, including those that are multidrug-resistant.

Variability of Accessory Proteins Rules the SARS-CoV-2 Pathogenicity (preprint)

The coronavirus disease 2019 (COVID-19) is caused by the Severe Acute Respiratory Syndrome Coronavirus-2 (SARS-CoV-2) which is pandemic with an estimated fatality rate less than 1% is ongoing. SARS-CoV-2 accessory proteins ORF3a, ORF6, ORF7a, ORF7b, ORF8, and ORF10 with putative functions to manipulate host immune mechanisms such as interferons, immune signaling receptor NLRP3 (NOD-, LRR-, and pyrin domain-containing 3) inflammasome, inflammatory cytokines such as interleukin {beta} (IL-1{beta}) are critical in COVID-19 pathology. Outspread variations of each of the six accessory proteins of all complete proteomes (available as of October 26, 2020, in the National Center for Biotechnology Information depository) of SARS-CoV-2, were observed across six continents. Across all continents, the decreasing order of percentage of unique variations in the accessory proteins was found to be ORF3a>ORF8>ORF7a>ORF6>ORF10>ORF7b. The highest and lowest unique variations of ORF3a were observed in South America and Oceania, respectively. This finding suggests that the wide variations of accessory proteins seem to govern the pathogenicity of SARS-CoV-2, and consequently, certain propositions and recommendations can be made in the public interest.

A zebrafish model for COVID-19 recapitulates olfactory and cardiovascular pathophysiologies caused by SARS-CoV-2 (preprint)

The COVID-19 pandemic has prompted the search for animal models that recapitulate the pathophysiology observed in humans infected with SARS-CoV-2 and allow rapid and high throughput testing of drugs and vaccines. Exposure of larvae to SARS-CoV-2 Spike (S) receptor binding domain (RBD) recombinant protein was sufficient to elevate larval heart rate and treatment with captopril, an ACE inhibitor, reverted this effect. Intranasal administration of SARS-CoV-2 S RBD in adult zebrafish recombinant protein caused severe olfactory and mild renal histopathology. Zebrafish intranasally treated with SARS-CoV-2 S RBD became hyposmic within minutes and completely anosmic by 1 day to a broad-spectrum of odorants including bile acids and food. Single cell RNA-Seq of the adult zebrafish olfactory organ indicated widespread loss of expression of olfactory receptors as well as inflammatory responses in sustentacular, endothelial, and myeloid cell clusters. Exposure of wildtype zebrafish larvae to SARS-CoV-2 in water did not support active viral replication but caused a sustained inhibition of ace2 expression, triggered type 1 cytokine responses and inhibited type 2 cytokine responses. Combined, our results establish adult and larval zebrafish as useful models to investigate pathophysiological effects of SARS-CoV-2 and perform pre-clinical drug testing and validation in an inexpensive, high throughput vertebrate model.

Forecasting COVID-19 cases at the Amazon region: a comparison of classical and machine learning models (preprint)

BACKGROUNDSince the first reports of COVID-19, decision-makers have been using traditional epidemiological models to predict the days to come. However, the enhancement of computational power, the demand for adaptable predictive frameworks, the short past of the disease, and uncertainties related to input data and prediction rules, also make other classical and machine learning techniques viable options. OBJECTIVEThis study investigates the efficiency of six models in forecasting COVID-19 confirmed cases with 17 days ahead. We compare the models autoregressive integrated moving average (ARIMA), Holt-Winters, support vector regression (SVR), k-nearest neighbors regressor (KNN), random trees regressor (RTR), seasonal linear regression with change-points (Prophet), and simple logistic regression (SLR). MATERIAL AND METHODSWe implement the models to data provided by the health surveillance secretary of Amapaa, a Brazilian state fully carved in the Amazon rainforest, which has been experiencing high infection rates. We evaluate the models according to their capacity to forecast in different historical scenarios of the COVID-19 progression, such as exponential increases, sudden decreases, and stability periods of daily cases. To do so, we use a rolling forward splitting approach for out-of-sample validation. We employ the metrics RMSE, R-squared, and sMAPE in evaluating the model in different cross-validation sections. FINDINGSAll models outperform SLG, especially Holt-Winters, that performs satisfactorily in all scenarios. SVR and ARIMA have better performances in isolated scenarios. To implement the comparisons, we have created a web application, which is available online. CONCLUSIONThis work represents an effort to assist the decision-makers of Amapa in future decisions to come, especially under scenarios of sudden variations in the number of confirmed cases of Amapa, which would be caused, for instance, by new contamination waves or vaccination. It is also an attempt to highlight alternative models that could be used in future epidemics.

Possible transmission flow of SARS-CoV-2 based on ACE2 features (preprint)

Angiotensin-converting enzyme 2 (ACE2) is the cellular receptor for the Severe Acute Respiratory Syndrome Coronavirus 2 (SARS-CoV-2) that is engendering the severe coronavirus disease 2019 (COVID-19) pandemic. The spike (S) protein receptor-binding domain (RBD) of SARS-CoV-2 binds to the three sub-domains viz. amino acids (aa) 22-42, aa 79-84, and aa 330-393 of ACE2 on human cells to initiate entry. It was reported earlier that the receptor utilization capacity of ACE2 proteins from different species, such as cats, chimpanzees, dogs, and cattle, are different. A comprehensive analysis of ACE2 receptors of nineteen species was carried out in this study, and the findings propose a possible SARS-CoV-2 transmission flow across these nineteen species.

SARS-CoV-2 infection induces mixed M1/M2 phenotype in circulating monocytes and alterations in both dendritic cell and monocyte subsets (preprint)

Clinical manifestations of SARS-CoV-2 infection range from mild to critically severe. The aim of the study was to highlight the immunological events associated with the severity of SARS-CoV-2 infection, with an emphasis on cells of innate immunity. Thirty COVID-19 patients with mild/moderate symptoms and 27 patients with severe/critically severe symptoms were recruited from the Clinical Center of Kragujevac during April 2020. Flow cytometric analysis was performed to reveal phenotypic and functional alterations of peripheral blood cells and to correlate them with the severity of the disease. In severe cases, the number of T and B lymphocytes, dendritic cells, NK cells, and HLA-DR-expressing cells was drastically decreased. In the monocyte population proportion between certain subsets was disturbed and cells coexpressing markers of M1 and M2 monocytes were found in intermediate and non-classical subsets. In mild cases decline in lymphocyte number was less pronounced and innate immunity was preserved as indicated by an increased number of myeloid and activated dendritic cells, NK cells that expressed activation marker at the same level as in control and by low expression of M2 marker in monocyte population. In patients with severe disease, both innate and adoptive immunity are devastated, while in patients with mild symptoms decline in lymphocyte number is lesser, and the innate immunity is preserved.