A Deep Learning Approach for Missing Data Imputation of Rating Scales Assessing Attention-Deficit Hyperactivity Disorder.

A variety of tools and methods have been used to measure behavioral symptoms of attention-deficit/hyperactivity disorder (ADHD). Missing data is a major concern in ADHD behavioral studies. This study used a deep learning method to impute missing data in ADHD rating scales and evaluated the ability of the imputed dataset (i.e., the imputed data replacing the original missing values) to distinguish youths with ADHD from youths without ADHD. The data were collected from 1220 youths, 799 of whom had an ADHD diagnosis, and 421 were typically developing (TD) youths without ADHD, recruited in Northern Taiwan. Participants were assessed using the Conners' Continuous Performance Test, the Chinese versions of the Conners' rating scale-revised: short form for parent and teacher reports, and the Swanson, Nolan, and Pelham, version IV scale for parent and teacher reports. We used deep learning, with information from the original complete dataset (referred to as the reference dataset), to perform missing data imputation and generate an imputation order according to the imputed accuracy of each question. We evaluated the effectiveness of imputation using support vector machine to classify the ADHD and TD groups in the imputed dataset. The imputed dataset can classify ADHD vs. TD up to 89% accuracy, which did not differ from the classification accuracy (89%) using the reference dataset. Most of the behaviors related to oppositional behaviors rated by teachers and hyperactivity/impulsivity rated by both parents and teachers showed high discriminatory accuracy to distinguish ADHD from non-ADHD. Our findings support a deep learning solution for missing data imputation without introducing bias to the data.

Bacterial Composition and Diversity in Breast Milk Samples from Mothers Living in Taiwan and Mainland China.

Human breast milk is widely recognized as the best source of nutrients for healthy growth and development of infants; it contains a diverse microbiota. Here, we characterized the diversity of the microbiota in the breast milk of East Asian women and assessed whether delivery mode influenced the microbiota in the milk of healthy breast-feeding mothers. We profiled the microbiota in breast milk samples collected from 133 healthy mothers in Taiwan and in six regions of mainland China (Central, East, North, Northeast, South, and Southwest China) by using 16S rRNA pyrosequencing. Lactation stage (months postpartum when the milk sample was collected) and maternal body mass index did not influence the breast milk microbiota. Bacterial composition at the family level differed significantly among samples from the seven geographical regions. The five most predominant bacterial families were Streptococcaceae (mean relative abundance: 24.4%), Pseudomonadaceae (14.0%), Staphylococcaceae (12.2%), Lactobacillaceae (6.2%), and Oxalobacteraceae (4.8%). The microbial profiles were classified into three clusters, driven by Staphylococcaceae (abundance in Cluster 1: 42.1%), Streptococcaceae (Cluster 2: 48.5%), or Pseudomonadaceae (Cluster 3: 26.5%). Microbial network analysis at the genus level revealed that the abundances of the Gram-positive Staphylococcus, Streptococcus, and Rothia were negatively correlated with those of the Gram-negative Acinetobacter, Bacteroides, Halomonas, Herbaspirillum, and Pseudomonas. Milk from mothers who had undergone Caesarian section (C-section group) had a significantly higher abundance of Lactobacillus (P < 0.05) and a higher number of unique unclassified operational taxonomic units (OTUs) (P < 0.001) than that from mothers who had undergone vaginal delivery (vaginal group). These findings revealed that (i) geographic differences in the microbial profiles were found in breast milk from mothers living in Taiwan and mainland China, (ii) the predominant bacterial families Streptococcaceae, Staphylococcaceae, and Pseudomonadaceae were key components for forming three respective clusters, and (iii) a significantly greater number of unique OTUs was found in the breast milk from mothers who had undergone C-section than from those who had delivered vaginally.

Comparative genomic analysis of bacteriocin-producing Weissella cibaria 110.

Weissella cibaria 110 was isolated from plaa-som, a Thai fermented fish product, and known to produce the weissellicin 110 bacteriocin. We carried out comprehensive comparative genomic analysis of W. cibaria 110 with four other non-bacteriocin-producing W. cibaria strains and identified potential antibiotic-resistant genes. We further identified a type III restriction-modification system, a TA system, and a bacteriocin gene cluster that are unique in W. cibaria 110. Genes related to bacteriocin biosynthesis are organized in clusters and are encoded with minimum genetic machinery consisting of structural cognate immunity genes, including ABC transporter and immunity protein. Finally, we predicted W. cibaria 110 to produce a class IId bacteriocin, weissellicin 110, which is 31 amino acids in length and contains a 21-amino-acid N-terminal leader peptide. This is the first bacteriocin-producing sequencing genome in W. cibaria, and we describe the difference between the bacteriocin-producing and non bacteriocin-producing strains from genome point of view.

Unique clustering genes in the bacterial chromosome affecting the type-III secretion of enterohaemorrhagic Escherichia coli.

Bioinformatics analysis was used to search for unknown genes that might influence the phenotypic presentations of enterohaemorrhagic Escherichia coli (EHEC). By so doing and using the known genomic data from EHEC O157 : H7 and K-12, it has been deduced that genes Z4863 to Z4866 of EHEC do not exist in K-12 strains. These four gene sequences have low degrees of homology (18-40 % amino acid identities) to a set of genes in K-12, which have been known to encode fatty acid biosynthesis enzymes. We referred these four consecutive genes as a fasyn cluster and found that deletion of fasyn from EHEC resulted in a defective type-III secretion (T3S). This deletion apparently did not decrease the amounts of the T3S proteins ectopically expressed from plasmids. Examination of the corresponding mRNAs by real-time PCR revealed that the mRNAs readily decreased in the fasyn-deleted mutant and this suppressive effect on the mRNA levels appeared to spread across all lee operons. Complementation with fasyn reverted the T3S-deficient phenotype. Furthermore, this reversion was also seen when the mutant was supplemented with locus of enterocyte effacement activators (Ler or GrlA). Thus, these unique clustering genes located apart from locus of enterocyte effacement on the bacterial chromosome also play a role in affecting T3S of EHEC.

Lessons from the Largest Epidemic of Avian Influenza Viruses in Taiwan, 2015.

The largest epidemic of avian influenza (AI) in history attacked poultry and wild birds throughout Taiwan starting January 6, 2015. This study analyzed surveillance results, epidemiologic characteristics, and viral sequences by using government-released information, with the intention to provide recommendations to minimize future pandemic influenza. The H5 clade 2.3.4.4 highly pathogenic AI viruses (HPAIVs) had not been detected in Taiwan before 2015. During this epidemic, four types of etiologic agents were identified: the three novel subtypes H5N2, H5N8, and H5N3 clade 2.3.4.4 HPAIVs and one endemic chicken H5N2 subtype (Mexican-like lineage) of low pathogenic AI viruses. Cocirculation of mixed subtypes also occurred, with H5N2 clade 2.3.4.4 HPAIVs accompanied by the H5N8 and H5N3 subtypes or old H5N2 viruses in the same farm. More than 90% of domestic geese died from this AI epidemic; geese were affected the most at the early outbreaks. The epidemic peaked in mid-January for all three novel H5 subtypes. Spatial epidemiology found that most affected areas were located in southwestern coastal areas. In terrestrial poultry (mostly chickens), different geographic distributions of AI virus subtypes were detected, with hot spots of H5N2 clade 2.3.4.4 vs. past-endemic old H5N2 viruses in Changhwa (P = 0.03) and Yunlin (P = 0.007) counties, respectively, of central Taiwan. Phylogenetic and sequence analyses of all the early 10 Taiwan H5 clade 2.3.4.4 isolates covering the three subtypes showed that they were very different from the HA of the past local H5 viruses from domestic ducks (75%-80%) and chickens (70%-75%). However, they had the highest sequence identity percentages (99.53%-100%), with the HA of A/crane/Kagoshima/KU13/2014(H5N8) isolated on December 7, 2014, in Japan being higher than those of recent American and Korean H5 HPAIVs [A/Northern pintail/Washington/40964/2014 (H5N2) and A/gyrfalcon/Washington/41088-6/2014 (H5N8): 99.02%-99.54% and A/Baikal teal/Korea/Donglim3/2014 (H5N8): 98.61%-99.08%], implying a likely common ancestor of these H5 clade 2.3.4.4 viruses. The multiple subtypes of H5 clade 2.3.4.4 HPAIVs imply high viral reassortment. We recommend establishing an integrated surveillance system, involving clinical, virologic, and serologic surveillance in poultry and wild birds, swine and other mammals prevalent on multiple-animal mixed-type traditional farms, and high-risk human populations, as a crucially important step to minimize future pandemic influenza.

Exploring comprehensive within-motif dependence of transcription factor binding in Escherichia coli.

Modeling the binding of transcription factors helps to decipher the control logic behind transcriptional regulatory networks. Position weight matrix is commonly used to describe a binding motif but assumes statistical independence between positions. Although current approaches take within-motif dependence into account for better predictive performance, these models usually rely on prior knowledge and incorporate simple positional dependence to describe binding motifs. The inability to take complex within-motif dependence into account may result in an incomplete representation of binding motifs. In this work, we applied association rule mining techniques and constructed models to explore within-motif dependence for transcription factors in Escherichia coli. Our models can reflect transcription factor-DNA recognition where the explored dependence correlates with the binding specificity. We also propose a graphical representation of the explored within-motif dependence to illustrate the final binding configurations. Understanding the binding configurations also enables us to fine-tune or design transcription factor binding sites, and we attempt to present the configurations through exploring within-motif dependence.

Genome architecture of Lactobacillus plantarum PS128, a probiotic strain with potential immunomodulatory activity.

BACKGROUND: Clinical and preclinical observations indicate that Lactobacillus plantarum has anti-inflammatory activity and may regulate the immune responses of its hosts when ingested. Recently, modification of teichoic acids (TAs) produced by L. plantarum was reported as a key to regulating the systemic immune response in mice. However, data linking TA-related genetic determinants and the immunomodulatory effect are limited. To provide genomic information for elucidating the underlying mechanism of immunomodulation by L. plantarum, we sequenced the genome of L. plantarum strain PS128. RESULTS: The PS128 genome contains 11 contigs (3,325,806 bp; 44.42% GC content) after hybrid assembly of sequences derived with Illumina MiSeq and PacBio RSII systems. The most abundant functions of the protein-coding genes are carbohydrate, amino acid, and protein metabolism. The 16S rDNA sequences of PS128 are closest to the sequences of L. plantarum WCFS1 and B21; these three strains form a distinct clade based on 16S rDNA sequences. PS128 shares core genes encoding the metabolism, transport, and modification of TAs with other sequenced L. plantarum strains. Compared with the TA-related genes of other completely sequenced L. plantarum strains, the PS128 contains more lipoteichoic acid exporter genes. CONCLUSIONS: We determined the draft genome sequence of PS128 and compared its TA-related genes with those of other L. plantarum strains. Shared genomic features with respect to TA-related subsystems may be important clues to the mechanism by which L. plantarum regulates its host immune responses, but unique TA-related genetic determinants should be further investigated to elucidate strain-specific immunomodulatory effects.

Behavior training reverses asymmetry in hippocampal transcriptome of the cav3.2 knockout mice.

Homozygous Cav3.2 knockout mice, which are defective in the pore-forming subunit of a low voltage activated T-type calcium channel, have been documented to show impaired maintenance of late-phase long-term potentiation (L-LTP) and defective retrieval of context-associated fear memory. To investigate the role of Cav3.2 in global gene expression, we performed a microarray transcriptome study on the hippocampi of the Cav3.2-/- mice and their wild-type littermates, either naïve (untrained) or trace fear conditioned. We found a significant left-right asymmetric effect on the hippocampal transcriptome caused by the Cav3.2 knockout. Between the naive Cav3.2-/- and the naive wild-type mice, 3522 differentially expressed genes (DEGs) were found in the left hippocampus, but only 4 DEGs were found in the right hippocampus. Remarkably, the effect of Cav3.2 knockout was partially reversed by trace fear conditioning. The number of DEGs in the left hippocampus was reduced to 6 in the Cav3.2 knockout mice after trace fear conditioning, compared with the wild-type naïve mice. To our knowledge, these results demonstrate for the first time the asymmetric effects of the Cav3.2 and its partial reversal by behavior training on the hippocampal transcriptome.

Hierarchical clustering of breast cancer methylomes revealed differentially methylated and expressed breast cancer genes.

Oncogenic transformation of normal cells often involves epigenetic alterations, including histone modification and DNA methylation. We conducted whole-genome bisulfite sequencing to determine the DNA methylomes of normal breast, fibroadenoma, invasive ductal carcinomas and MCF7. The emergence, disappearance, expansion and contraction of kilobase-sized hypomethylated regions (HMRs) and the hypomethylation of the megabase-sized partially methylated domains (PMDs) are the major forms of methylation changes observed in breast tumor samples. Hierarchical clustering of HMR revealed tumor-specific hypermethylated clusters and differential methylated enhancers specific to normal or breast cancer cell lines. Joint analysis of gene expression and DNA methylation data of normal breast and breast cancer cells identified differentially methylated and expressed genes associated with breast and/or ovarian cancers in cancer-specific HMR clusters. Furthermore, aberrant patterns of X-chromosome inactivation (XCI) was found in breast cancer cell lines as well as breast tumor samples in the TCGA BRCA (breast invasive carcinoma) dataset. They were characterized with differentially hypermethylated XIST promoter, reduced expression of XIST, and over-expression of hypomethylated X-linked genes. High expressions of these genes were significantly associated with lower survival rates in breast cancer patients. Comprehensive analysis of the normal and breast tumor methylomes suggests selective targeting of DNA methylation changes during breast cancer progression. The weak causal relationship between DNA methylation and gene expression observed in this study is evident of more complex role of DNA methylation in the regulation of gene expression in human epigenetics that deserves further investigation.

A17, the First Sequenced Strain of Lactococcus lactis subsp. cremoris with Potential Immunomodulatory Functions.

Lactococcus lactis subsp. cremoris A17, isolated from Taiwan fermented cabbage, is the first sequenced strain of L. lactis subsp. cremoris with immunomodulatory activity and antiallergic functions. The resulting A17 draft genome contains 2,679,936 bp and indicates that A17 is a potential exopolysaccharide-producing strain without any known virulence gene.

Diagnostics for stochastic genome-scale modeling via model slicing and debugging.

Modeling of biological behavior has evolved from simple gene expression plots represented by mathematical equations to genome-scale systems biology networks. However, due to obstacles in complexity and scalability of creating genome-scale models, several biological modelers have turned to programming or scripting languages and away from modeling fundamentals. In doing so, they have traded the ability to have exchangeable, standardized model representation formats, while those that remain true to standardized model representation are faced with challenges in model complexity and analysis. We have developed a model diagnostic methodology inspired by program slicing and debugging and demonstrate the effectiveness of the methodology on a genome-scale metabolic network model published in the BioModels database. The computer-aided identification revealed specific points of interest such as reversibility of reactions, initialization of species amounts, and parameter estimation that improved a candidate cell's adenosine triphosphate production. We then compared the advantages of our methodology over other modeling techniques such as model checking and model reduction. A software application that implements the methodology is available at http://gel.ym.edu.tw/gcs/.

Draft Genome Sequence of Lactobacillus pobuzihii E100301T.

Lactobacillus pobuzihii E100301(T) is a novel Lactobacillus species previously isolated from pobuzihi (fermented cummingcordia) in Taiwan. Phylogenetically, this strain is closest to Lactobacillus acidipiscis, but its phenotypic characteristics can be clearly distinguished from those of L. acidipiscis. We present the draft genome sequence of strain L. pobuzihii E100301(T).

Microbial diversity analysis of fermented mung beans (Lu-Doh-Huang) by using pyrosequencing and culture methods.

In Taiwanese alternative medicine Lu-doh-huang (also called Pracparatum mungo), mung beans are mixed with various herbal medicines and undergo a 4-stage process of anaerobic fermentation. Here we used high-throughput sequencing of the 16S rRNA gene to profile the bacterial community structure of Lu-doh-huang samples. Pyrosequencing of samples obtained at 7 points during fermentation revealed 9 phyla, 264 genera, and 586 species of bacteria. While mung beans were inside bamboo sections (stages 1 and 2 of the fermentation process), family Lactobacillaceae and genus Lactobacillus emerged in highest abundance; Lactobacillus plantarum was broadly distributed among these samples. During stage 3, the bacterial distribution shifted to family Porphyromonadaceae, and Butyricimonas virosa became the predominant microbial component. Thereafter, bacterial counts decreased dramatically, and organisms were too few to be detected during stage 4. In addition, the microbial compositions of the liquids used for soaking bamboo sections were dramatically different: Exiguobacterium mexicanum predominated in the fermented soybean solution whereas B. virosa was predominant in running spring water. Furthermore, our results from pyrosequencing paralleled those we obtained by using the traditional culture method, which targets lactic acid bacteria. In conclusion, the microbial communities during Lu-doh-huang fermentation were markedly diverse, and pyrosequencing revealed a complete picture of the microbial consortium.

Genome-wide PhoB binding and gene expression profiles reveal the hierarchical gene regulatory network of phosphate starvation in Escherichia coli.

The phosphate starvation response in bacteria has been studied extensively for the past few decades and the phosphate-limiting signal is known to be mediated via the PhoBR two-component system. However, the global DNA binding profile of the response regulator PhoB and the PhoB downstream responses are currently unclear. In this study, chromatin immunoprecipitation for PhoB was combined with high-density tiling array (ChIP-chip) as well as gene expression microarray to reveal the first global down-stream responses of the responding regulator, PhoB in E. coli. Based on our ChIP-chip experimental data, forty-three binding sites were identified throughout the genome and the known PhoB binding pattern was updated by identifying the conserved pattern from these sites. From the gene expression microarray data analysis, 287 differentially expressed genes were identified in the presence of PhoB activity. By comparing the results obtained from our ChIP-chip and microarray experiments, we were also able to identify genes that were directly or indirectly affected through PhoB regulation. Nineteen out of these 287 differentially expressed genes were identified as the genes directly regulated by PhoB. Seven of the 19 directly regulated genes (including phoB) are transcriptional regulators. These transcriptional regulators then further pass the signal of phosphate starvation down to the remaining differentially expressed genes. Our results unveiled the genome-wide binding profile of PhoB and the downstream responses under phosphate starvation. We also present the hierarchical structure of the phosphate sensing regulatory network. The data suggest that PhoB plays protective roles in membrane integrity and oxidative stress reduction during phosphate starvation.

Protein domain repetition is enriched in Streptococcal cell-surface proteins.

Tandem repetition of domain in protein sequence occurs in all three domains of life. It creates protein diversity and adds functional complexity in organisms. In this work, we analyzed 52 streptococcal genomes and found 3748 proteins contained domain repeats. Proteins not harboring domain repeats are significantly enriched in cytoplasm, whereas proteins with domain repeats are significantly enriched in cytoplasmic membrane, cell wall and extracellular locations. Domain repetition occurs most frequently in S. pneumoniae and least in S. thermophilus and S. pyogenes. DUF1542 is the highest repeated domain in a single protein, followed by Rib, CW_binding_1, G5 and HemolysinCabind. 3D structures of 24 repeat-containing proteins were predicted to investigate the structural and functional effect of domain repetition. Several repeat-containing streptococcal cell surface proteins are known to be virulence-associated. Surface-associated tandem domain-containing proteins without experimental functional characterization may be potentially involved in the pathogenesis of streptococci and deserve further investigation.

Highly conserved influenza A virus epitope sequences as candidates of H3N2 flu vaccine targets.

This study focused on identifying the conserved epitopes in a single subtype A (H3N2)-as candidates for vaccine targets. We identified a total of 32 conserved epitopes in four viral proteins [22 HA, 4PB1, 3 NA, 3 NP]. Evaluation of conserved epitopes in coverage during 1968-2010 revealed that (1) 12 HA conserved epitopes were highly present in the circulating viruses; (2) the remaining 10 HA conserved epitopes appeared with lower percentage but a significantly increasing trend after 1989 [p<0.001]; and (3) the conserved epitopes in NA, NP and PB1 are also highly frequent in wild-type viruses. These conserved epitopes also covered an extremely high percentage of the 16 vaccine strains during the 42 year period. The identification of highly conserved epitopes using our approach can also be applied to develop broad-spectrum vaccines.

Proteomics characterization of cytoplasmic and lipid-associated membrane proteins of human pathogen Mycoplasma fermentans M64.

Mycoplasma fermentans is a potent human pathogen which has been implicated in several diseases. Notably, its lipid-associated membrane proteins (LAMPs) play a role in immunomodulation and development of infection-associated inflammatory diseases. However, the systematic protein identification of pathogenic M. fermentans has not been reported. From our recent sequencing results of M. fermentans M64 isolated from human respiratory tract, its genome is around 1.1 Mb and encodes 1050 predicted protein-coding genes. In the present study, soluble proteome of M. fermentans was resolved and analyzed using two-dimensional gel electrophoresis. In addition, Triton X-114 extraction was carried out to enrich amphiphilic proteins including putative lipoproteins and membrane proteins. Subsequent mass spectrometric analyses of these proteins had identified a total of 181 M. fermentans ORFs. Further bioinformatics analysis of these ORFs encoding proteins with known or so far unknown orthologues among bacteria revealed that a total of 131 proteins are homologous to known proteins, 11 proteins are conserved hypothetical proteins, and the remaining 39 proteins are likely M. fermentans-specific proteins. Moreover, Triton X-114-enriched fraction was shown to activate NF-kB activity of raw264.7 macrophage and a total of 21 lipoproteins with predicted signal peptide were identified therefrom. Together, our work provides the first proteome reference map of M. fermentans as well as several putative virulence-associated proteins as diagnostic markers or vaccine candidates for further functional study of this human pathogen.

CAGO: a software tool for dynamic visual comparison and correlation measurement of genome organization.

CAGO (Comparative Analysis of Genome Organization) is developed to address two critical shortcomings of conventional genome atlas plotters: lack of dynamic exploratory functions and absence of signal analysis for genomic properties. With dynamic exploratory functions, users can directly manipulate chromosome tracks of a genome atlas and intuitively identify distinct genomic signals by visual comparison. Signal analysis of genomic properties can further detect inconspicuous patterns from noisy genomic properties and calculate correlations between genomic properties across various genomes. To implement dynamic exploratory functions, CAGO presents each genome atlas in Scalable Vector Graphics (SVG) format and allows users to interact with it using a SVG viewer through JavaScript. Signal analysis functions are implemented using R statistical software and a discrete wavelet transformation package waveslim. CAGO is not only a plotter for generating complex genome atlases, but also a platform for exploring genome atlases with dynamic exploratory functions for visual comparison and with signal analysis for comparing genomic properties across multiple organisms. The web-based application of CAGO, its source code, user guides, video demos, and live examples are publicly available and can be accessed at http://cbs.ym.edu.tw/cago.

Sequencing and comparative genome analysis of two pathogenic Streptococcus gallolyticus subspecies: genome plasticity, adaptation and virulence.

Streptococcus gallolyticus infections in humans are often associated with bacteremia, infective endocarditis and colon cancers. The disease manifestations are different depending on the subspecies of S. gallolyticus causing the infection. Here, we present the complete genomes of S. gallolyticus ATCC 43143 (biotype I) and S. pasteurianus ATCC 43144 (biotype II.2). The genomic differences between the two biotypes were characterized with comparative genomic analyses. The chromosome of ATCC 43143 and ATCC 43144 are 2,36 and 2,10 Mb in length and encode 2246 and 1869 CDS respectively. The organization and genomic contents of both genomes were most similar to the recently published S. gallolyticus UCN34, where 2073 (92%) and 1607 (86%) of the ATCC 43143 and ATCC 43144 CDS were conserved in UCN34 respectively. There are around 600 CDS conserved in all Streptococcus genomes, indicating the Streptococcus genus has a small core-genome (constitute around 30% of total CDS) and substantial evolutionary plasticity. We identified eight and five regions of genome plasticity in ATCC 43143 and ATCC 43144 respectively. Within these regions, several proteins were recognized to contribute to the fitness and virulence of each of the two subspecies. We have also predicted putative cell-surface associated proteins that could play a role in adherence to host tissues, leading to persistent infections causing sub-acute and chronic diseases in humans. This study showed evidence that the S. gallolyticus still possesses genes making it suitable in a rumen environment, whereas the ability for S. pasteurianus to live in rumen is reduced. The genome heterogeneity and genetic diversity among the two biotypes, especially membrane and lipoproteins, most likely contribute to the differences in the pathogenesis of the two S. gallolyticus biotypes and the type of disease an infected patient eventually develops.

T4-Like genome organization of the Escherichia coli O157:H7 lytic phage AR1.

We report the genome organization and analysis of the first completely sequenced T4-like phage, AR1, of Escherichia coli O157:H7. Unlike most of the other sequenced phages of O157:H7, which belong to the temperate Podoviridae and Siphoviridae families, AR1 is a T4-like phage known to efficiently infect this pathogenic bacterial strain. The 167,435-bp AR1 genome is currently the largest among all the sequenced E. coli O157:H7 phages. It carries a total of 281 potential open reading frames (ORFs) and 10 putative tRNA genes. Of these, 126 predicted proteins could be classified into six viral orthologous group categories, with at least 18 proteins of the structural protein category having been detected by tandem mass spectrometry. Comparative genomic analysis of AR1 and four other completely sequenced T4-like genomes (RB32, RB69, T4, and JS98) indicated that they share a well-organized and highly conserved core genome, particularly in the regions encoding DNA replication and virion structural proteins. The major diverse features between these phages include the modules of distal tail fibers and the types and numbers of internal proteins, tRNA genes, and mobile elements. Codon usage analysis suggested that the presence of AR1-encoded tRNAs may be relevant to the codon usage of structural proteins. Furthermore, protein sequence analysis of AR1 gp37, a potential receptor binding protein, indicated that eight residues in the C terminus are unique to O157:H7 T4-like phages AR1 and PP01. These residues are known to be located in the T4 receptor recognition domain, and they may contribute to specificity for adsorption to the O157:H7 strain.