Coding genomes with gapped pattern graph convolutional network.

MOTIVATION: Genome sequencing technologies reveal a huge amount of genomic sequences. Neural network-based methods can be prime candidates for retrieving insights from these sequences because of their applicability to large and diverse datasets. However, the highly variable lengths of genome sequences severely impair the presentation of sequences as input to the neural network. Genetic variations further complicate tasks that involve sequence comparison or alignment. RESULTS: Inspired by the theory and applications of "spaced seeds," we propose a graph representation of genome sequences called "gapped pattern graph." These graphs can be transformed through a Graph Convolutional Network to form lower-dimensional embeddings for downstream tasks. On the basis of the gapped pattern graphs, we implemented a neural network model and demonstrated its performance on diverse tasks involving microbe and mammalian genome data. Our method consistently outperformed all the other state-of-the-art methods across various metrics on all tasks, especially for the sequences with limited homology to the training data. In addition, our model was able to identify distinct gapped pattern signatures from the sequences. AVAILABILITY AND IMPLEMENTATION: The framework is available at https://github.com/deepomicslab/GCNFrame.

Fatal swine acute diarrhoea syndrome caused by an HKU2-related coronavirus of bat origin.

Cross-species transmission of viruses from wildlife animal reservoirs poses a marked threat to human and animal health 1 . Bats have been recognized as one of the most important reservoirs for emerging viruses and the transmission of a coronavirus that originated in bats to humans via intermediate hosts was responsible for the high-impact emerging zoonosis, severe acute respiratory syndrome (SARS) 2-10 . Here we provide virological, epidemiological, evolutionary and experimental evidence that a novel HKU2-related bat coronavirus, swine acute diarrhoea syndrome coronavirus (SADS-CoV), is the aetiological agent that was responsible for a large-scale outbreak of fatal disease in pigs in China that has caused the death of 24,693 piglets across four farms. Notably, the outbreak began in Guangdong province in the vicinity of the origin of the SARS pandemic. Furthermore, we identified SADS-related CoVs with 96-98% sequence identity in 9.8% (58 out of 591) of anal swabs collected from bats in Guangdong province during 2013-2016, predominantly in horseshoe bats (Rhinolophus spp.) that are known reservoirs of SARS-related CoVs. We found that there were striking similarities between the SADS and SARS outbreaks in geographical, temporal, ecological and aetiological settings. This study highlights the importance of identifying coronavirus diversity and distribution in bats to mitigate future outbreaks that could threaten livestock, public health and economic growth.

On triangle inequalities of correlation-based distances for gene expression profiles.

BACKGROUND: Distance functions are fundamental for evaluating the differences between gene expression profiles. Such a function would output a low value if the profiles are strongly correlated-either negatively or positively-and vice versa. One popular distance function is the absolute correlation distance, [Formula: see text], where [Formula: see text] is similarity measure, such as Pearson or Spearman correlation. However, the absolute correlation distance fails to fulfill the triangle inequality, which would have guaranteed better performance at vector quantization, allowed fast data localization, as well as accelerated data clustering. RESULTS: In this work, we propose [Formula: see text] as an alternative. We prove that [Formula: see text] satisfies the triangle inequality when [Formula: see text] represents Pearson correlation, Spearman correlation, or Cosine similarity. We show [Formula: see text] to be better than [Formula: see text], another variant of [Formula: see text] that satisfies the triangle inequality, both analytically as well as experimentally. We empirically compared [Formula: see text] with [Formula: see text] in gene clustering and sample clustering experiment by real-world biological data. The two distances performed similarly in both gene clustering and sample clustering in hierarchical clustering and PAM (partitioning around medoids) clustering. However, [Formula: see text] demonstrated more robust clustering. According to the bootstrap experiment, [Formula: see text] generated more robust sample pair partition more frequently (P-value [Formula: see text]). The statistics on the time a class "dissolved" also support the advantage of [Formula: see text] in robustness. CONCLUSION: [Formula: see text], as a variant of absolute correlation distance, satisfies the triangle inequality and is capable for more robust clustering.

Deep learning enables de novo peptide sequencing from data-independent-acquisition mass spectrometry.

We present DeepNovo-DIA, a de novo peptide-sequencing method for data-independent acquisition (DIA) mass spectrometry data. We use neural networks to capture precursor and fragment ions across m/z, retention-time, and intensity dimensions. They are then further integrated with peptide sequence patterns to address the problem of highly multiplexed spectra. DIA coupled with de novo sequencing allowed us to identify novel peptides in human antibodies and antigens.

Characterization and Genomic Analysis of Bacteriophage vB_KpnM_IME346 Targeting Clinical Klebsiella pneumoniae Strain of the K63 Capsular Type.

A Klebsiella pneumoniae bacteriophage (vB_KpnM_IME346) was isolated from a hospital sewage sample. This bacteriophage specifically infects a clinical K. pneumoniae strain with a K63 capsular polysaccharide structure. The phage genome was evaluated by next-generation sequencing, which revealed a linear double-stranded DNA genome consisting of 49,482 base pairs with a G+C content of 49.1%. The latent period of vB_KpnM_IME346 was shown to be 20 min, and the burst size was 25-30 pfu (plaque-forming units)/infected cell. Transmission electron microscopy and phylogenetic analysis showed that the JD001-like phage belongs to the genus Jedunavirus of the family Myoviridae. The newly isolated vB_KpnM_IME346 shows infectivity in the clinical host K. pneumoniae KP576 strain, indicating that it is a promising alternative to antibacterial agents for removing K. pneumoniae from patients.

Genetic diversity and evolutionary dynamics of Ebola virus in Sierra Leone.

A novel Ebola virus (EBOV) first identified in March 2014 has infected more than 25,000 people in West Africa, resulting in more than 10,000 deaths. Preliminary analyses of genome sequences of 81 EBOV collected from March to June 2014 from Guinea and Sierra Leone suggest that the 2014 EBOV originated from an independent transmission event from its natural reservoir followed by sustained human-to-human infections. It has been reported that the EBOV genome variation might have an effect on the efficacy of sequence-based virus detection and candidate therapeutics. However, only limited viral information has been available since July 2014, when the outbreak entered a rapid growth phase. Here we describe 175 full-length EBOV genome sequences from five severely stricken districts in Sierra Leone from 28 September to 11 November 2014. We found that the 2014 EBOV has become more phylogenetically and genetically diverse from July to November 2014, characterized by the emergence of multiple novel lineages. The substitution rate for the 2014 EBOV was estimated to be 1.23 × 10(-3) substitutions per site per year (95% highest posterior density interval, 1.04 × 10(-3) to 1.41 × 10(-3) substitutions per site per year), approximating to that observed between previous EBOV outbreaks. The sharp increase in genetic diversity of the 2014 EBOV warrants extensive EBOV surveillance in Sierra Leone, Guinea and Liberia to better understand the viral evolution and transmission dynamics of the ongoing outbreak. These data will facilitate the international efforts to develop vaccines and therapeutics.

Characteristics and complete genome sequence of the virulent Vibrio alginolyticus phage VAP7, isolated in Hainan, China.

A novel Vibrio alginolyticus phage, VAP7, was isolated from seawater collected from Sanya, Hainan province, China. Whole-genome sequencing analysis revealed that phage VAP7 has a linear, double-stranded DNA genome of 144,685 bp with an average G+C content of 41.9% and a high degree of sequence similarity to Vibrio phage VP-1. Annotation results identified 193 open reading frames and one transfer RNA-encoding gene in the phage genome. The morphology and the results of phylogenetic analysis suggest that VAP7 should be classified as a new member of the family Ackermannviridae. Moreover, phage VAP7 grew over a wide pH (5.0-10.0) and temperature (4-40 °C) range. Host-range experiments revealed that VAP7 could infect 31 Vibrio alginolyticus strains. Thus, VAP7 infecting Vibrio alginolyticus strains represents a potential new candidate for use in phage therapy.

Complete genome sequence of a novel bovine hepacivirus from Yunnan, China.

We detected a novel bovine hepacivirus N (HNV) subtype, IME_BovHep_01, in the serum of cattle in Tengchong, Yunnan, China, by high-throughput sequencing. The complete genome of IME_BovHep_01, was sequenced using an Illumina MiSeq sequencer and found to be 8850 nt in length, encoding one hypothetical protein. BLASTn analysis showed that the genome sequence shared similarity with the bovine hepacivirus isolate BovHepV_209/Ger/2014, with 88% query coverage and 70.8% identity. However, the highest similarity was to bovine hepacivirus N strain BRBovHep_RS963, for which only a partial genome sequence is available, with 68% query coverage and 81.5% identity. Sequence comparisons and phylogenetic analysis suggested that IME_BovHep_01 is a novel HNV subtype. Importantly, IME_BovHep_01 is the first member of this new genotype for which the complete genome sequence was determined.

A novel freshwater cyanophage vB_MelS-Me-ZS1 infecting bloom-forming cyanobacterium Microcystis elabens.

Blooms of cyanobacteria cause enormous losses in both the economy and environment. Cyanophages are of great potential for fighting blooming cyanobacteria. Research report on cyanophage of bloom-forming cyanobacterium, Microcystis elabens is deficient. vB_MelS-Me-ZS1 (abbreviated as Me-ZS1) was isolated from fresh water by double-layer agar plate method using M. elabens. TEM exhibited that cyanosiphovirus Me-ZS1 has an icosahedral head about 60 nm in diameter, and a noncontractile tail approximately 260 nm. Experimental infection against 15 cyanobacterial strains showed that Me-ZS1 can infect 12 strains across taxonomic orders (Chroococcales, Nostocales and Oscillatoriales). High-throughput sequencing and bioinformatics analysis revealed that Me-ZS1 has a double-stranded DNA genome of 49,665 bp, with a G + C content of 58.22%, and 73 predicted open reading frames (ORFs). BLASTn and ORF comparisons showed that Me-ZS1 shares very low homology with the public sequences, and the phylogenetic tree based on TerL indicated that Me-ZS1 may delegate a novel and genetically distinct clade of Siphoviridae phages. In microcosm experiment, Me-ZS1 represented apparent effect on reducing relative abundance of cyanobacteria, increasing relative abundance of Saprospiraceae and protecting brocade carp (Carassius auratus) in cyanobacterial bloom water. This study isolated and characterized a novel broad-host-range Microcystis phage Me-ZS1 presenting a genetically distinct clade of freshwater cyanophage. The features of cyanophage Me-ZS1 provide a potential solution to the loss caused by cyanobacterial bloom.

Isolation and Characterization of a Novel Bacteriophage Infecting Carbapenem-Resistant Klebsiella pneumoniae.

A novel virulent phage, vB_KpnP_IME337, isolated from a hospital sewage in Beijing, China, that infects carbapenem-resistant Klebsiella pneumoniae KN2 capsular type was identified and characterized. Next-generation sequencing and genome analysis revealed that vB_KpnP_IME337 had a linear double-stranded genome with a length of 44,266 base pairs and G+C content of 53.7%. Fifty-two putative open reading frames were identified, and no transfer RNA-encoding genes were detected. BLASTn analysis revealed that phage vB_KpnP_IME337 had the highest sequence similarity with Klebsiella phage phiBO1E, with genome coverage of 79%. Based on morphology, phage vB_KpnP_IME337 was determined to belong to the family Podoviridae of the order Caudovirales. It was shown that phage vB_KpnP_IME337 had an infection duration of ~ 90 min and 10 min latent period, and a highly specific to host strain. In conclusion, phage vB_KpnP_IME337 may be a promising alternative candidate to antibiotic treatment for controlling diseases caused by drug-resistant K. pneumoniae.

De novo peptide sequencing by deep learning.

De novo peptide sequencing from tandem MS data is the key technology in proteomics for the characterization of proteins, especially for new sequences, such as mAbs. In this study, we propose a deep neural network model, DeepNovo, for de novo peptide sequencing. DeepNovo architecture combines recent advances in convolutional neural networks and recurrent neural networks to learn features of tandem mass spectra, fragment ions, and sequence patterns of peptides. The networks are further integrated with local dynamic programming to solve the complex optimization task of de novo sequencing. We evaluated the method on a wide variety of species and found that DeepNovo considerably outperformed state of the art methods, achieving 7.7-22.9% higher accuracy at the amino acid level and 38.1-64.0% higher accuracy at the peptide level. We further used DeepNovo to automatically reconstruct the complete sequences of antibody light and heavy chains of mouse, achieving 97.5-100% coverage and 97.2-99.5% accuracy, without assisting databases. Moreover, DeepNovo is retrainable to adapt to any sources of data and provides a complete end-to-end training and prediction solution to the de novo sequencing problem. Not only does our study extend the deep learning revolution to a new field, but it also shows an innovative approach in solving optimization problems by using deep learning and dynamic programming.

Complete genome analysis of a novel temperate bacteriophage induced from Corynebacterium striatum.

A temperate bacteriophage, IME1320_01, was induced by mitomycin C treatment from Corynebacterium striatum. This phage possesses a double-stranded DNA genome of 40,086 bp with a G+C content of 58%. A total of 53 putative open reading frames (ORFs) were identified in its genome. BLASTn analysis revealed that IME1320_01 had the highest sequence similarity to Corynebacterium striatum strain 216, with a genome homology coverage of 44% and highest sequence identity of 95%. The termini of the phage genome was non-redundant, with a 13-nt 3'-protruding cohesive end. To the best of our knowledge, phage IME1320_01 is the first inducible phage to be identified in Corynebacterium striatum.

Identification of a lytic Pseudomonas aeruginosa phage depolymerase and its anti-biofilm effect and bactericidal contribution to serum.

Pseudomonas aeruginosa (P. aeruginosa) infection has imposed a great threat to patients with cystic fibrosis. With the emergence of multidrug-resistant P. aeruginosa, developing an alternative anti-microbial strategy is indispensable and more urgent than ever. In this study, a lytic P. aeruginosa phage was isolated from the sewage of a hospital, and one protein was predicted as the depolymerase-like protein by genomic sequence analysis, it includes two catalytic regions, the Pectate lyase_3 super family and Glycosyl hydrolase_28 super family. Further analysis demonstrated that recombinant depolymerase-like protein degraded P. aeruginosa exopolysaccharide and enhanced bactericidal activity mediated by serum in vitro. Additionally, this protein disrupted host bacterial biofilms. All of these results showed that the phage-derived depolymerase-like protein has the potential to be developed into an anti-microbial agent that targets P. aeruginosa.

Genetic diversity, virulence factors and farm-to-table spread pattern of Vibrio parahaemolyticus food-associated isolates.

Vibrio parahaemolyticus is the leading bacterial cause of seafood-associated gastroenteritis worldwide. Moreover, infections and outbreaks caused by V. parahaemolyticus has kept increasing over the last two decades. In this study, we investigated the genetic diversity, virulence factors and farm-to-table spread pattern of V. parahaemolyticus by analyzing 383 genomes of food-associated isolates. These strains were isolated from diverse sample types from six provinces of China in 2014, being classified into three tiers of the farm-to-table spread process: food production, circulation and consumption. The genetic diversity of V. parahaemolyticus in different classifications, including geographical location, sample type, source and spread tier, was similar, as the median number of pairwise SNPs within each classification was between 33,013 and 33,659. Specifically, there was no clear boundaries in genetic diversity of the isolates from inland vs. coastal provinces, as well as of those from freshwater vs. seawater products. Moreover, the virulence genes and genomic islands were only found in a small number of isolates, indicating a low disease risk of the food-associated isolates in this study. By further exploring 28 recently emerged clonal groups, we identified seven farm-to-table spread events, showing a common pattern of single-source radial spread accompanied with occasional gene gain/loss events. Generally speaking, our work highlighted the colonization of V. parahaemolyticus in inland provinces and freshwater environment, and provided a snapshot of the farm-to-table spread pattern of V. parahaemolyticus food-associated isolates. Our results showed the feasibility of tracking the farm-to-table spread of foodborne pathogen, which would help construct the whole genome sequencing-based molecular tracking network in the future.

Deep Omics.

Deep learning has revolutionized research in image processing, speech recognition, natural language processing, game playing, and will soon revolutionize research in proteomics and genomics. Through three examples in genomics, protein structure prediction, and proteomics, we demonstrate that deep learning is changing bioinformatics research, shifting from algorithm-centric to data-centric approaches.

First complete genome sequence of a virulent bacteriophage infecting the opportunistic pathogen Serratia rubidaea.

A Serratia rubidaea phage, vB_Sru IME250, was isolated from hospital sewage. The morphology suggested that phage vB_Sru IME250 should be classified as a member of the family Myoviridae. High-throughput sequencing revealed that the phage genome has 154,938 nucleotides and consists of 193 coding DNA sequences, 90 of which have putative functions. The genome of vB_Sru IME250 is a linear, double-stranded DNA with an average GC content of 40.04%. The phage has a relatively high similarity to Klebsiella phage 0507-KN2-1, with a genome coverage of 84%.

Complete genome sequence of a novel, virulent Ahjdlikevirus bacteriophage that infects Enterococcus faecium.

A novel virulent bacteriophage named vB_EfaP_IME199 that specifically infects Enterococcus faecium was isolated and characterized. Its optimal multiplicity of infection was 0.01, and it had a 30 minute outbreak period. High-throughput sequencing revealed that the phage has a dsDNA genome of 18,838 bp with 22 open reading frames. The genome has very low homology to all other bacteriophage sequences in the GenBank database. Run-off sequencing experiments confirmed that vB_EfaP_IME199 has short inverted terminal repeats. Phylogenetic analysis indicated that vB_EfaP_IME199 can be taxonomically classified as a new member of the genus Ahjdlikevirus of family Podoviridae.

Complete genome sequence of Menghai flavivirus, a novel insect-specific flavivirus from China.

Menghai flavivirus (MFV) was isolated from Aedes albopictus in Menghai county of Yunnan Province, China, during an arboviruses screening program in August 2010. Whole genome sequencing of MFV was performed using an Ion PGM™ Sequencer. The complete genome of MFV was 10897 nucleotides in length and encoded a polyprotein and fairly interesting flavivirus orf (FIFO). The polyprotein contained three flavivirus structural proteins (C, prM/M and E) and seven nonstructural proteins. Nucleotide BLAST analysis revealed that the MFV genome showed highest similarity to Xishuangbanna Aedes flavivirus, a novel insect-specific flavivirus recently isolated from the same area. These species shared a query cover of 99%, but only 71% identity, while FIFO showed no similarity with any of the published sequences. Genomic and phylogenetic analyses suggested that MFV was a novel species of the genus Flavivirus. Our findings enrich our understanding of the genetics and prevalence of the family Flaviviridae.

Complete genome sequence of Menghai rhabdovirus, a novel mosquito-borne rhabdovirus from China.

Menghai rhabdovirus (MRV) was isolated from Aedes albopictus in Menghai county of Yunnan Province, China, in August 2010. Whole-genome sequencing of MRV was performed using an Ion PGM™ Sequencer. We found that MRV is a single-stranded, negative-sense RNA virus. The complete genome of MRV has 10,744 nt, with short inverted repeat termini, encoding five typical rhabdovirus proteins (N, P, M, G, and L) and an additional small hypothetical protein. Nucleotide BLAST analysis using the BLASTn method showed that the genome sequence most similar to that of MRV is that of Arboretum virus (NC_025393.1), with a Max score of 322, query coverage of 14%, and 66% identity. Genomic and phylogenetic analyses both demonstrated that MRV should be considered a member of a novel species of the family Rhabdoviridae.