Ebola virus (EBOV) has caused several outbreaks as the consequence of spillover events from zoonotic sources and has resulted in huge death tolls. In spite of considerable progress, a thorough know-how regarding EBOV adaptation in various host species and detailed information about the potential reservoirs of EBOV still remains obscure. The present study was executed to examine the patterns of codon usage and its associated influence in the adaptation of EBOV to potential hosts that dwell in Africa, the origin of the viral outbreaks. Correspondence analysis (CA) revealed that the codon usage signature in EBOV is a complex interplay of factors including compositional bias and natural selection, with the latter having a more pronounced impact. Low codon usage bias in EBOV indicates a flexibility of the viruses in adapting to diverse range of hosts with different codon usage architectures. EBOV adaptation in potential hosts, as estimated by codon adaptation index (CAI) and relative codon deoptimization index (RCDI), revealed that the viruses were relatively better adapted to African primates than other mammals examined, which might account for the high fatality rate of primates owing to EBOV infection. Bats have been speculated as natural reservoirs of EBOV. In the present analysis it was interesting to note that EBOV displayed lower degrees of adaptation, as estimated by CAI and RCDI, with bats in comparison to the primate hosts. Lower degrees of adaptation might contribute to long-term coexistence and circulation of the viral pathogens in bat populations. Codon usage patterns of EBOV isolates associated with different outbreaks varied significantly, with discrete patterns between the West and Central African isolates. Additional evolutionary analyses indicated that the West African Epidemic began with an initial spillover infection and there was more than one population of EBOV circulating in the natural reservoir in the Democratic Republic of the Congo. The present study yields valuable information regarding the possible circulation of EBOV in various African mammals.
The H1N1/pdm2009 virus is a new triple-reassortant virus. While Eurasian avian-like and triple-reassortant swine influenza viruses are the direct ancestors of H1N1/pdm2009, the classic swine influenza virus facilitate the spectrum of influenza A diversity in pig population when the reassortant events occurred during 1998 to April 2009. The factors that facilitate the final formation of this gene constellation for H1N1/pdm2009 virus from this complex gene pool remain unknown. Since a novel successful virus should efficiently replicate and transmit in their hosts, in this study, we estimated the adaptability of the codon usage patterns of the pool of genes from these lineages of swine influenza viruses to the human expression system. We found that the MP and NA genes of Eurasian avian-like swine influenza viruses, and the PB2, PB1 and PA genes of triple-reassortant swine influenza viruses were best adapted to the human codon usage pattern. As these genes participated in the development of H1N1/pdm2009, they might help in viral replication and strengthen its competitiveness during its emergence. After its emergence in the human population, a gradual optimization of codon usage patterns between 2009 and 2019 to the human codon usage for the H1N1/pdm2009 genes was detected. This reveals that ongoing adaptive evolution, after its original incursion, occurred to further increase the adaptability of overall gene cassette to human expression system.
Object detection is essential in Computer Vision and is widely applied in all areas. This paper proposes a method called BAFPN. BAFPN is a new bidirectional Feature Pyramid Network that constructs accurate object detection networks based on YOLOv4 by implementing Adaptively Spatial Feature Fusion. Besides, Exponential Moving Average is used to improve the network performance. The developed network not only maintains high computing speed but also enhances the mAP by 4.3% when testing with the MS COCO dataset and when comparing it to the original YOLOv4. To further improve the performance, the trained model was pruned using the Batch Normalization layer's scaling factor, achieving an 18% reduction in size and improving object detection speed.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.