Coronavirus disease 2019 (COVID-19) has caused thousands of deaths worldwide and has become an urgent public health concern. The extraordinary interhuman transmission of this disease has urged scientists to examine the various facets of its pathogenic agent, the severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2). Herein, based on publicly available genomic data, we analyzed the codon usage co-adaptation profiles of SARS-CoV-2 and other respiratory coronaviruses (CoVs) with their human host, identified CoV-responsive human genes and their functional roles on the basis of both the relative synonymous codon usage (RSCU)-based correlation of viral genes with human genes and differential gene expression analysis, and predicted potential drugs for COVID-19 treatment based on these genes. The relatively high codon adaptation index (CAI) values (> 0.70) signposted the gene expressivity efficiency of CoVs in human. The ENc-GC3 plot indicated that SARS-CoV-2 genome was under strict selection pressure while SARS-CoV and MERS-CoV were under selection and mutational pressures. The RSCU-based correlation analysis indicated that the viral genomes shared similar codons with a panoply of human genes. The merging of RSCU-based correlation data and SARS-CoV-2-responsive differentially expressed genes allowed the identification of human genes potentially affected by SARS-CoV-2 infection. Functional enrichment analysis indicated that these genes were enriched in biological processes and pathways related to host response to viral infection and immune response. Using the drug-gene interaction database, we screened a list of drugs that could target these genes as potential COVID-19 therapeutics. Our findings not only will contribute in vaccine development but also provide a useful set of drugs that could guide practitioners in strategical monitoring of COVID-19. We recommend practitioners to scrupulously screen this list of predicted drugs in order to authenticate those qualified for treating COVID-19 symptoms.

Molecular mechanisms of the non-structural protein 1 (NS1) in influenza A-induced pathological changes remain ambiguous. This study explored the pathogenesis of human infection by influenza A viruses (IAVs) through identifying human genes with codon usage bias (CUB) similar to NS1 gene of these viruses based on the relative synonymous codon usage (RSCU). CUB of the IAV subtypes H1N1, H3N2, H3N8, H5N1, H5N2, H5N8, H7N9 and H9N2 was analyzed and the correlation of RSCU values of NS1 sequences with those of the human genes was calculated. The CUB of NS1 was uneven and codons ending with A/U were preferred. The ENC-GC3 and neutrality plots suggested natural selection as the main determinant for CUB. The RCDI, CAI and SiD values showed that the viruses had a high degree of adaptability to human. A total of 2155 human genes showed significant RSCU-based correlation (p < 0.05 and r > 0.5) with NS1 coding sequences and was considered as human genes with CUB similar to NS1 gene of IAV subtypes. Differences and similarities in the subtype-specific human protein–protein interaction (PPI) networks and their functions were recorded among IAVs subtypes, indicating that NS1 of each IAV subtype has a specific pathogenic mechanism. Processes and pathways involved in influenza, transcription, immune response and cell cycle were enriched in human gene sets retrieved based on the CUB of NS1 gene of IAV subtypes. The present work may advance our understanding on the mechanism of NS1 in human infections of IAV subtypes and shed light on the therapeutic options. Supplementary Information The online version contains supplementary material available at 10.1007/s10709-022-00155-9.

No abstract

No abstract

A masquerader is an attacker who gains illegitimate access to a user's account. Masquerade detection is one of the key problems of intrusion detection systems. Deep learning models that obtained state-of-the-art results in masquerade detection have failed to exhibit very high detection performance when data samples contain limited information. Alternatively, computationally cheaper and more memoryefficient traditional machine learning models suffer from less robust features, which hinders them in achieving high detection performance. The contributions of this paper are as follows: we introduce new features of variable-length UNIX command sequences (i.e., weighted occurrence frequencies of different orders) and integrate these features into an extended Markov-chain-based variable-length model. The detection performance of our model is evaluated on three publicly available and free datasets: Schonlau (SEA), Purdue (PU), and Greenberg. The results demonstrate that our model significantly improves the true positive rate (TPR), false positive rate, receiver operator characteristic, and threshold variance compared to the baselines (other Markov-chain-based variable-length models). Furthermore, in terms of the TPR, the proposed method is superior to a state-of-the-art deep learning model that uses a convolutional neural network on the PU and Greenberg datasets and a state-of-the-art sequence-alignment-hidden Markov model on the SEA dataset. Moreover, the proposed method is much more lightweight than the state-of-the-art models in terms of computational and memory complexity, and thus more suitable for real-time masquerade detection.