ARGNet: using deep neural networks for robust identification and classification of antibiotic resistance genes from sequences.

BACKGROUND: Emergence of antibiotic resistance in bacteria is an important threat to global health. Antibiotic resistance genes (ARGs) are some of the key components to define bacterial resistance and their spread in different environments. Identification of ARGs, particularly from high-throughput sequencing data of the specimens, is the state-of-the-art method for comprehensively monitoring their spread and evolution. Current computational methods to identify ARGs mainly rely on alignment-based sequence similarities with known ARGs. Such approaches are limited by choice of reference databases and may potentially miss novel ARGs. The similarity thresholds are usually simple and could not accommodate variations across different gene families and regions. It is also difficult to scale up when sequence data are increasing. RESULTS: In this study, we developed ARGNet, a deep neural network that incorporates an unsupervised learning autoencoder model to identify ARGs and a multiclass classification convolutional neural network to classify ARGs that do not depend on sequence alignment. This approach enables a more efficient discovery of both known and novel ARGs. ARGNet accepts both amino acid and nucleotide sequences of variable lengths, from partial (30-50 aa; 100-150 nt) sequences to full-length protein or genes, allowing its application in both target sequencing and metagenomic sequencing. Our performance evaluation showed that ARGNet outperformed other deep learning models including DeepARG and HMD-ARG in most of the application scenarios especially quasi-negative test and the analysis of prediction consistency with phylogenetic tree. ARGNet has a reduced inference runtime by up to 57% relative to DeepARG. CONCLUSIONS: ARGNet is flexible, efficient, and accurate at predicting a broad range of ARGs from the sequencing data. ARGNet is freely available at https://github.com/id-bioinfo/ARGNet , with an online service provided at https://ARGNet.hku.hk . Video Abstract.

Association between Gut Microbiota Composition and Long-Term Vaccine Immunogenicity following Three Doses of CoronaVac.

BACKGROUND: Neutralizing antibody level wanes with time after COVID-19 vaccination. We aimed to study the relationship between baseline gut microbiota and immunogenicity after three doses of CoronaVac. METHODS: This was a prospective cohort study recruiting three-dose CoronaVac recipients from two centers in Hong Kong. Blood samples were collected at baseline and one year post-first dose for virus microneutralization (vMN) assays to determine neutralization titers. The primary outcome was high immune response (defined as with vMN titer ≥ 40). Shotgun DNA metagenomic sequencing of baseline fecal samples identified potential bacterial species and metabolic pathways using Linear Discriminant Analysis Effect Size (LEfSe) analysis. Univariate and multivariable logistic regression models were used to identify high response predictors. RESULTS: In total, 36 subjects were recruited (median age: 52.7 years [IQR: 47.9-56.4]; male: 14 [38.9%]), and 18 had low immune response at one year post-first dose vaccination. Eubacterium rectale (log10LDA score = 4.15, p = 0.001; relative abundance of 1.4% vs. 0, p = 0.002), Collinsella aerofaciens (log10LDA score = 3.31, p = 0.037; 0.39% vs. 0.18%, p = 0.038), and Streptococcus salivarius (log10LDA score = 2.79, p = 0.021; 0.05% vs. 0.02%, p = 0.022) were enriched in low responders. The aOR of high immune response with E. rectale, C. aerofaciens, and S. salivarius was 0.03 (95% CI: 9.56 × 10-4-0.32), 0.03 (95% CI: 4.47 × 10-4-0.59), and 10.19 (95% CI: 0.81-323.88), respectively. S. salivarius had a positive correlation with pathways enriched in high responders like incomplete reductive TCA cycle (log10LDA score = 2.23). C. aerofaciens similarly correlated with amino acid biosynthesis-related pathways. These pathways all showed anti-inflammation functions. CONCLUSION: E. rectale,C. aerofaciens, and S. salivarius correlated with poorer long-term immunogenicity following three doses of CoronaVac.

The Predictive Value of Gut Microbiota Composition for Sustained Immunogenicity following Two Doses of CoronaVac.

CoronaVac immunogenicity decreases with time, and we aimed to investigate whether gut microbiota associate with longer-term immunogenicity of CoronaVac. This was a prospective cohort study recruiting two-dose CoronaVac recipients from three centres in Hong Kong. We collected blood samples at baseline and day 180 after the first dose and used chemiluminescence immunoassay to test for neutralizing antibodies (NAbs) against the receptor-binding domain (RBD) of wild-type SARS-CoV-2 virus. We performed shotgun metagenomic sequencing performed on baseline stool samples. The primary outcome was the NAb seroconversion rate (seropositivity defined as NAb ≥ 15AU/mL) at day 180. Linear discriminant analysis [LDA] effect size analysis was used to identify putative bacterial species and metabolic pathways. A univariate logistic regression model was used to derive the odds ratio (OR) of seropositivity with bacterial species. Of 119 CoronaVac recipients (median age: 53.4 years [IQR: 47.8-61.3]; male: 39 [32.8%]), only 8 (6.7%) remained seropositive at 6 months after vaccination. Bacteroides uniformis (log10LDA score = 4.39) and Bacteroides eggerthii (log10LDA score = 3.89) were significantly enriched in seropositive than seronegative participants. Seropositivity was associated with B. eggerthii (OR: 5.73; 95% CI: 1.32-29.55; p = 0.022) and B. uniformis with borderline significance (OR: 3.27; 95% CI: 0.73-14.72; p = 0.110). Additionally, B. uniformis was positively correlated with most enriched metabolic pathways in seropositive vaccinees, including the superpathway of adenosine nucleotide de novo biosynthesis I (log10LDA score = 2.88) and II (log10LDA score = 2.91), as well as pathways related to vitamin B biosynthesis, all of which are known to promote immune functions. In conclusion, certain gut bacterial species (B. eggerthii and B. uniformis) and metabolic pathways were associated with longer-term CoronaVac immunogenicity.

Surveillance of Rodent Pests for SARS-CoV-2 and Other Coronaviruses, Hong Kong.

We report surveillance conducted in 217 pestiferous rodents in Hong Kong for severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2). We did not detect SARS-CoV-2 RNA but identified 1 seropositive rodent, suggesting exposure to a virus antigenically similar to SARS-CoV-2. Potential exposure of urban rodents to SARS-CoV-2 cannot be ruled out.

Sequencing and characterisation of complete mitogenome DNA for Rasbora sarawakensis (Cypriniformes: Cyprinidae: Rasbora) with phylogenetic consideration.

The Blueline Rasbora (Rasbora sarawakensis) is a small ray-finned fish categorized under the genus Rasbora in the Cyprinidae family. In this study, the complete mitogenome sequence of R. sarawakensis was sequenced using four primers targeting overlapping regions. The mitogenome is 16,709 bp in size, accommodating 22 transfer RNA genes, 13 protein-coding genes, two ribosomal RNA genes and a putative control region. Identical gene organisation was detected between this species and other genus counterparts. The heavy strand houses 28 genes while the light strand stores the other nine genes. Most protein-coding genes employ ATG as start codon, excluding COI gene, which utilizes GTG instead. The central conserved sequence blocks (CSB-F, CSB-E and CSB-D), variable sequence blocks (CSB-3, CSB-2 and CSB-1) as well as the terminal associated sequence (TAS) are conserved in the control region. The maximum likelihood phylogenetic tree revealed the divergence of R. sarawakensis from the basal region of the Rasbora clade, where its evolutionary relationships with R. maculatus and R. pauciperforata are poorly resolved as indicated by the low bootstrap values. This work acts as steppingstone towards further molecular evolution and population genetics studies of Rasbora genus in future.

Sequencing and Characterisation of Complete Mitochondrial DNA Genome for Trigonopoma pauciperforatum (Cypriniformes: Cyprinidae: Danioninae) with Phylogenetic Consideration.

The Trigonopoma pauciperforatum or the redstripe rasbora is a cyprinid commonly found in marshes and swampy areas with slight acidic tannin-stained water in the tropics. In this study, the complete mitogenome sequence of T. pauciperforatum was first amplified in two parts using two pairs of overlapping primers and then sequenced. The size of the mitogenome is 16,707 bp, encompassing 22 transfer RNA genes, 13 protein-coding genes, two ribosomal RNA genes and a putative control region. Identical gene organisation was detected between this species and other family members. The heavy strand accommodates 28 genes while the light strand houses the remaining nine genes. Most protein-coding genes utilise ATG as start codon except for COI gene which uses GTG instead. The terminal associated sequence (TAS), central conserved sequence block (CSB-F, CSB-D and CSB-E) as well as variable sequence block (CSB-1, CSB-2 and CSB-3) are conserved in the control region. The maximum likelihood phylogenetic tree revealed the divergence of T. pauciperforatum from the basal region of the major clade, where its evolutionary relationships with Boraras maculatus, Rasbora cephalotaenia and R. daniconius are poorly resolved as suggested by the low bootstrap values. This work contributes towards the genetic resource enrichment for peat swamp conservation and comprehensive in-depth comparisons across other phylogenetic researches done on the Rasbora-related genus.

Sequencing and characterization of complete mitogenome DNA of Rasbora tornieri (Cypriniformes: Cyprinidae: Rasbora) and its evolutionary significance.

The yellowtail rasbora (Rasbora tornieri) is a miniature ray-finned fish categorized under the genus Rasbora in the family of Cyprinidae. In this study, a complete mitogenome sequence of R. tornieri was sequenced using four primers targeting two halves of the mitogenome with overlapping flanking regions. The size of mitogenome was 16,573 bp, housing 22 transfer RNA genes, 13 protein-coding genes, two ribosomal RNA genes and a putative control region. Identical gene organization was detected between this species and other members of Rasbora genus. The heavy strand encompassed 28 genes while the light strand accommodated the other nine genes. Most protein-coding genes execute ATG as start codon, excluding COI and ND3 genes, which utilized GTG instead. The central conserved sequence blocks (CSB-E, CSB-F and CSB-D), variable sequence blocks (CSB-1, CSB-3 and CSB-2) as well as the terminal associated sequence (TAS) were conserved within the control region. The maximum likelihood phylogenetic family tree revealed the divergence of R. tornieri from the basal region of the Rasbora clade, where its evolutionary relationships with other Rasbora members are poorly resolved as indicated by the low bootstrap values. This work acts as window for further population genetics and molecular evolution studies of Rasbora genus in future.

Identifying SARS-CoV-2-related coronaviruses in Malayan pangolins.

The ongoing outbreak of viral pneumonia in China and across the world is associated with a new coronavirus, SARS-CoV-21. This outbreak has been tentatively associated with a seafood market in Wuhan, China, where the sale of wild animals may be the source of zoonotic infection2. Although bats are probable reservoir hosts for SARS-CoV-2, the identity of any intermediate host that may have facilitated transfer to humans is unknown. Here we report the identification of SARS-CoV-2-related coronaviruses in Malayan pangolins (Manis javanica) seized in anti-smuggling operations in southern China. Metagenomic sequencing identified pangolin-associated coronaviruses that belong to two sub-lineages of SARS-CoV-2-related coronaviruses, including one that exhibits strong similarity in the receptor-binding domain to SARS-CoV-2. The discovery of multiple lineages of pangolin coronavirus and their similarity to SARS-CoV-2 suggests that pangolins should be considered as possible hosts in the emergence of new coronaviruses and should be removed from wet markets to prevent zoonotic transmission.

Systematic analysis of the caridean shrimp superfamily Pandaloidea (Crustacea: Decapoda) based on molecular and morphological evidence.

One of the systematically controversial superfamilies in Caridea is the predominately deep-sea or cold water Pandaloidea, largely because this species-rich group of nearly 200 species in 25 genera exhibits a very high diversity of body forms and ecology. Although the relationships amongst the taxa within Pandaloidea have been repeatedly discussed based on morphology, no comprehensive molecular phylogeny exists. In this study, we present the first molecular phylogeny of the group, based on a combined dataset of two mitochondrial (12S and 16S rRNA) and six nuclear (ATP synthase ß-subunit, enolase, glyceraldehyde-3-phosphate dehydrogenase, histone 3, phosphoenolpyruvate carboxykinase and sodium-potassium ATPase α-subunit) markers, based on 62 species (about 1/3 of known biodiversity) in 22 genera (88% of genera) of two pandaloid families (Pandalidae, Thalassocarididae) and outgroups from seven other caridean families. With generally high support, the relationships within the clade are fully resolved. Pandalidae is shown to be paraphyletic with Thalassocarididae deeply nested within as a monophyletic group, and the latter is herein considered to be a synonym of Pandalidae. Five major clades are recovered, with the shallow water genera Anachlorocurtis, Chlorocurtis, Chlorotocella and Miropandalus forming a sister clade to the remaining genera. At the genus level, the phylogeny indicates Plesionika, Heterocarpus and Pandalus to be not monophyletic. The validity of Pandalopsis, Stylopandalus and Calipandalus is challenged and these genera are considered herein to be junior synonyms of Pandalus (Pandalopsis) and Plesionika (Stylopandalus and Calipandalus). Although not fully resolved, some evidence potentially considers Nothocaris to be a valid genus. Ancestral State Reconstruction successfully recovered 15 synapomorphies for the major clades, with 11 of them reported to be of systematic significance for the first time.

Molecular phylogeny of Pasiphaeidae (Crustacea, Decapoda, Caridea) reveals systematic incongruence of the current classification.

Caridean shrimps constitute one of the most diverse groups of decapod crustaceans, notwithstanding their poorly resolved infraordinal relationships. One of the systematically controversial families in Caridea is the predominantly pelagic Pasiphaeidae, comprises 101 species in seven genera. Pasiphaeidae species exhibit high morphological disparity, as well as ecological niche width, inhabiting shallow to very deep waters (>4000m). The present work presents the first molecular phylogeny of the family, based on a combined dataset of six mitochondrial and nuclear gene markers (12S rDNA, 16S rDNA, histone 3, sodium-potassium ATPase α-subunit, enolase and ATP synthase ß-subunit) from 33 species belonged to six genera of Pasiphaeidae with 19 species from 12 other caridean families as outgroup taxa. Maximum likelihood and Bayesian inference analyses conducted on the concatenated dataset of 2265bp suggest the family Pasiphaeidae is not monophyletic, with Psathyrocaris more closely related to other carideans than to the other five pasiphaeid genera included in this analysis. Leptochela occupies a sister position to the remaining genera and is genetically quite distant from them. At the generic level, the analysis supports the monophyly of Pasiphaea, Leptochela and Psathyrocaris, while Eupasiphae is shown to be paraphyletic, closely related to Parapasiphae and Glyphus. The present molecular result strongly implies that certain morphological characters used in the present systematic delineation within Pasiphaeidae may not be synapomorphies and the classification within the family needs to be urgently revised.