Reconstruction of the genome-scale metabolic network model of Sinorhizobium fredii CCBAU45436 for free-living and symbiotic states.

Sinorhizobium fredii CCBAU45436 is an excellent rhizobium that plays an important role in agricultural production. However, there still needs more comprehensive understanding of the metabolic system of S. fredii CCBAU45436, which hinders its application in agriculture. Therefore, based on the first-generation metabolic model iCC541 we developed a new genome-scale metabolic model iAQY970, which contains 970 genes, 1,052 reactions, 942 metabolites and is scored 89% in the MEMOTE test. Cell growth phenotype predicted by iAQY970 is 81.7% consistent with the experimental data. The results of mapping the proteome data under free-living and symbiosis conditions to the model showed that the biomass production rate in the logarithmic phase was faster than that in the stable phase, and the nitrogen fixation efficiency of rhizobia parasitized in cultivated soybean was higher than that in wild-type soybean, which was consistent with the actual situation. In the symbiotic condition, there are 184 genes that would affect growth, of which 94 are essential; In the free-living condition, there are 143 genes that influence growth, of which 78 are essential. Among them, 86 of the 94 essential genes in the symbiotic condition were consistent with the prediction of iCC541, and 44 essential genes were confirmed by literature information; meanwhile, 30 genes were identified by DEG and 33 genes were identified by Geptop. In addition, we extracted four key nitrogen fixation modules from the model and predicted that sulfite reductase (EC 1.8.7.1) and nitrogenase (EC 1.18.6.1) as the target enzymes to enhance nitrogen fixation by MOMA, which provided a potential focus for strain optimization. Through the comprehensive metabolic model, we can better understand the metabolic capabilities of S. fredii CCBAU45436 and make full use of it in the future.

GMMAD: a comprehensive database of human gut microbial metabolite associations with diseases.

BACKGROUND: The natural products, metabolites, of gut microbes are crucial effect factors on diseases. Comprehensive identification and annotation of relationships among disease, metabolites, and microbes can provide efficient and targeted solutions towards understanding the mechanism of complex disease and development of new markers and drugs. RESULTS: We developed Gut Microbial Metabolite Association with Disease (GMMAD), a manually curated database of associations among human diseases, gut microbes, and metabolites of gut microbes. Here, this initial release (i) contains 3,836 disease-microbe associations and 879,263 microbe-metabolite associations, which were extracted from literatures and available resources and then experienced our manual curation; (ii) defines an association strength score and a confidence score. With these two scores, GMMAD predicted 220,690 disease-metabolite associations, where the metabolites all belong to the gut microbes. We think that the positive effective (with both scores higher than suggested thresholds) associations will help identify disease marker and understand the pathogenic mechanism from the sense of gut microbes. The negative effective associations would be taken as biomarkers and have the potential as drug candidates. Literature proofs supported our proposal with experimental consistence; (iii) provides a user-friendly web interface that allows users to browse, search, and download information on associations among diseases, metabolites, and microbes. The resource is freely available at http://guolab.whu.edu.cn/GMMAD . CONCLUSIONS: As the online-available unique resource for gut microbial metabolite-disease associations, GMMAD is helpful for researchers to explore mechanisms of disease- metabolite-microbe and screen the drug and marker candidates for different diseases.

Mslar: Microbial synthetic lethal and rescue database.

Synthetic lethality (SL) occurs when mutations in two genes together lead to cell or organism death, while a single mutation in either gene does not have a significant impact. This concept can also be extended to three or more genes for SL. Computational and experimental methods have been developed to predict and verify SL gene pairs, especially for yeast and Escherichia coli. However, there is currently a lack of a specialized platform to collect microbial SL gene pairs. Therefore, we designed a synthetic interaction database for microbial genetics that collects 13,313 SL and 2,994 Synthetic Rescue (SR) gene pairs that are reported in the literature, as well as 86,981 putative SL pairs got through homologous transfer method in 281 bacterial genomes. Our database website provides multiple functions such as search, browse, visualization, and Blast. Based on the SL interaction data in the S. cerevisiae, we review the issue of duplications' essentiality and observed that the duplicated genes and singletons have a similar ratio of being essential when we consider both individual and SL. The Microbial Synthetic Lethal and Rescue Database (Mslar) is expected to be a useful reference resource for researchers interested in the SL and SR genes of microorganisms. Mslar is open freely to everyone and available on the web at http://guolab.whu.edu.cn/Mslar/.

Prediction of protein-coding small ORFs in multi-species using integrated sequence-derived features and the random forest model.

Proteins encoded by small open reading frames (sORFs) can serve as functional elements playing important roles in vivo. Such sORFs also constitute the potential pool for facilitating the de novo gene birth, driving evolutionary innovation and species diversity. Therefore, their theoretical and experimental identification has become a critical issue. Herein, we proposed a protein-coding sORFs prediction method merely based on integrative sequence-derived features. Our prediction performance is better or comparable compared with other nine prevalent methods, which shows that our method can provide a relatively reliable research tool for the prediction of protein-coding sORFs. Our method allows users to estimate the potential expression of a queried sORF, which has been demonstrated by the correlation analysis between our possibility estimation and codon adaption index (CAI). Based on the features that we used, we demonstrated that the sequence features of the protein-coding sORFs in the two domains have significant differences implying that it might be a relatively hard task in terms of cross-domain prediction, hence domain-specific models were developed, which allowed users to predict protein-coding sORFs both in eukaryotes and prokaryotes. Finally, a web-server was developed and provided to boost and facilitate the study of the related field, which is freely available at http://guolab.whu.edu.cn/codingCapacity/index.html.

Discovery of Drug Candidates for Specific Human Disease Based on Natural Products of Gut Microbes.

The beneficial metabolites of the microbiome could be used as a tool for screening drugs that have the potential for the therapy of various human diseases. Narrowing down the range of beneficial metabolite candidates in specific diseases was primarily a key step for further validation in model organisms. Herein, we proposed a reasonable hypothesis that the metabolites existing commonly in multiple beneficial (or negatively associated) bacteria might have a high probability of being effective drug candidates for specific diseases. According to this hypothesis, we screened metabolites associated with seven human diseases. For type I diabetes, 45 out of 88 screened metabolites had been reported as potential drugs in the literature. Meanwhile, 18 of these metabolites were specific to type I diabetes. Additionally, metabolite correlation could reflect disease relationships in some sense. Our results have demonstrated the potential of bioinformatics mining gut microbes' metabolites as drug candidates based on reported numerous microbe-disease associations and the Virtual Metabolic Human database. More subtle methods would be developed to ensure more accurate predictions.

EcoliGD: An Online Tool for Designing Escherichia coli Genome.

Synthetic biology is an important interdisciplinary field that has emerged in this century, focusing on the rewriting and reprogramming of DNA through the cycles of "design-edit", and so, the cell's own operating system, its genome, is naturally coming into focus. Here, we propose EcoliGD, an online genome design tool with a visual interactive interface and the function of browsing information, as well as the ability to perform insertion, exchange, deletion, and codon replacement operations on the E. coli genome and display the results in real-time. Users can utilize EcoliGD to check various functional characteristic about E. coli genes, to help them build their genomes. Furthermore, we also collected experimentally verified large genomic segments that have been successfully deleted from the genome for users to choose from and simplify the genome. EcoliGD can help recode the entire E. coli genome, providing a novel way to explore the diversity and function of this microorganism. The EcoliGD web tool is available at http://guolab.whu.edu.cn/EcoliGD/.

T-G-A Deficiency Pattern in Protein-Coding Genes and Its Potential Reason.

If a stop codon appears within one gene, then its translation will be terminated earlier than expected. False folding of premature protein will be adverse to the host; hence, all functional genes would tend to avoid the intragenic stop codons. Therefore, we hypothesize that there will be less frequency of nucleotides corresponding to stop codons at each codon position of genes. Here, we validate this inference by investigating the nucleotide frequency at a large scale and results from 19,911 prokaryote genomes revealed that nucleotides coinciding with stop codons indeed have the lowest frequency in most genomes. Interestingly, genes with three types of stop codons all tend to follow a T-G-A deficiency pattern, suggesting that the property of avoiding intragenic termination pressure is the same and the major stop codon TGA plays a dominant role in this effect. Finally, a positive correlation between the TGA deficiency extent and the base length was observed in start-experimentally verified genes of Escherichia coli (E. coli). This strengthens the proof of our hypothesis. The T-G-A deficiency pattern observed would help to understand the evolution of codon usage tactics in extant organisms.

Anti-CRISPRdb v2.2: an online repository of anti-CRISPR proteins including information on inhibitory mechanisms, activities and neighbors of curated anti-CRISPR proteins.

We previously released the Anti-CRISPRdb database hosting anti-CRISPR proteins (Acrs) and associated information. Since then, the number of known Acr families, types, structures and inhibitory activities has accumulated over time, and Acr neighbors can be used as a candidate pool for screening Acrs in further studies. Therefore, we here updated the database to include the new available information. Our newly updated database shows several improvements: (i) it comprises more entries and families because it includes both Acrs reported in the most recent literatures and Acrs obtained via performing homologous alignment; (ii) the prediction of Acr neighbors is integrated into Anti-CRISPRdb v2.2, and users can identify novel Acrs from these candidates; and (iii) this version includes experimental information on the inhibitory strength and stage for Acr-Cas/Acr-CRISPR pairs, motivating the development of tools for predicting specific inhibitory abilities. Additionally, a parameter, the rank of codon usage bias (CUBRank), was proposed and provided in the new version, which showed a positive relationship with predicted result from AcRanker; hence, it can be used as an indicator for proteins to be Acrs. CUBRank can be used to estimate the possibility of genes occurring within genome island-a hotspot hosting potential genes encoding Acrs. Based on CUBRank and Anti-CRISPRdb, we also gave the first glimpse for the emergence of Acr genes (acrs). DATABASE URL: http://guolab.whu.edu.cn/anti-CRISPRdb.

Geptop 2.0: Accurately Select Essential Genes from the List of Protein-Coding Genes in Prokaryotic Genomes.

Computational tool composites alternative way to identify essential genes and it is low-cost and time-efficient. Based on experimental essentiality sets deposited in the databases DEG and OGEE as reference, we developed an automatically computational tool named Geptop to select essential genes from the set of protein-coding genes in a prokaryotic genome, which utilizes the strategy of reciprocally best hit for homology search and evolutionary distance for weight assigning. The latest version of Geptop is 2.0 ( http://guolab.whu.edu.cn/geptop ), which can predict gene essentiality with the mean AUC 0f 0.84 in prokaryotes and is more stable. The chapter is to briefly introduce the tool and tell how to use it.

Consistent Clustering Pattern of Prokaryotic Genes Based on Base Frequency at the Second Codon Position and its Association with Functional Category Preference.

In 2002, our research group observed a gene clustering pattern based on the base frequency of A versus T at the second codon position in the genome of Vibrio cholera and found that the functional category distribution of genes in the two clusters was different. With the availability of a large number of sequenced genomes, we performed a systematic investigation of A2-T2 distribution and found that 2694 out of 2764 prokaryotic genomes have an optimal clustering number of two, indicating a consistent pattern. Analysis of the functional categories of the coding genes in each cluster in 1483 prokaryotic genomes indicated, that 99.33% of the genomes exhibited a significant difference (p < 0.01) in function distribution between the two clusters. Specifically, functional category P was overrepresented in the small cluster of 98.65% of genomes, whereas categories J, K, and L were overrepresented in the larger cluster of over 98.52% of genomes. Lineage analysis uncovered that these preferences appear consistently across all phyla. Overall, our work revealed an almost universal clustering pattern based on the relative frequency of A2 versus T2 and its role in functional category preference. These findings will promote the understanding of the rationality of theoretical prediction of functional classes of genes from their nucleotide sequences and how protein function is determined by DNA sequence.

Quantitative elucidation of associations between nucleotide identity and physicochemical properties of amino acids and the functional insight.

Studies on codon property would deepen our understanding of the origin of primitive life and enlighten biotechnical application. Here, we proposed a quantitative measurement of codon-amino acid association and found that seven out of 13 physicochemical properties have stronger associations with the nucleotide identity at the second codon position, indicating that protein structure and function may associate more closely with it than the other two sites. When extending the effect of codon-amino acid association to protein level, it was found that the correlation between the second codon position (measured by the relative frequencies of nucleobase T and A at this codon site) and hydrophobicity (by the form of GRAVY value) became stronger with 96% genomes having R > 0.90 and p < 1e-60. Furthermore, we revealed that informational genes encoding proteins have lower GRAVY values than operational proteins (p < 3e-37) in both prokaryotic and eukaryotic genomes. The above results reveal a complete link from codon identity (A2 versus T2) to amino acid property (hydrophilic versus hydrophobic) and then to protein functions (informational versus operational). Hence, our work may help to understand how the nucleotide sequence determines protein function.

BioMaster: An Integrated Database and Analytic Platform to Provide Comprehensive Information About BioBrick Parts.

Synthetic biology seeks to create new biological parts, devices, and systems, and to reconfigure existing natural biological systems for custom-designed purposes. The standardized BioBrick parts are the foundation of synthetic biology. The incomplete and flawed metadata of BioBrick parts, however, are a major obstacle for designing genetic circuit easily, quickly, and accurately. Here, a database termed BioMaster http://www.biomaster-uestc.cn was developed to extensively complement information about BioBrick parts, which includes 47,934 items of BioBrick parts from the international Genetically Engineered Machine (iGEM) Registry with more comprehensive information integrated from 10 databases, providing corresponding information about functions, activities, interactions, and related literature. Moreover, BioMaster is also a user-friendly platform for retrieval and analyses of relevant information on BioBrick parts.

CEG 2.0: an updated database of clusters of essential genes including eukaryotic organisms.

Essential genes are key elements for organisms to maintain their living. Building databases that store essential genes in the form of homologous clusters, rather than storing them as a singleton, can provide more enlightening information such as the general essentiality of homologous genes in multiple organisms. In 2013, the first database to store prokaryotic essential genes in clusters, CEG (Clusters of Essential Genes), was constructed. Afterward, the amount of available data for essential genes increased by a factor >3 since the last revision. Herein, we updated CEG to version 2, including more prokaryotic essential genes (from 16 gene datasets to 29 gene datasets) and newly added eukaryotic essential genes (nine species), specifically the human essential genes of 12 cancer cell lines. For prokaryotes, information associated with drug targets, such as protein structure, ligand-protein interaction, virulence factor and matched drugs, is also provided. Finally, we provided the service of essential gene prediction for both prokaryotes and eukaryotes. We hope our updated database will benefit more researchers in drug targets and evolutionary genomics. Database URL: http://cefg.uestc.cn/ceg.

Comprehensive review of the identification of essential genes using computational methods: focusing on feature implementation and assessment.

Essential genes have attracted increasing attention in recent years due to the important functions of these genes in organisms. Among the methods used to identify the essential genes, accurate and efficient computational methods can make up for the deficiencies of expensive and time-consuming experimental technologies. In this review, we have collected researches on essential gene predictions in prokaryotes and eukaryotes and summarized the five predominant types of features used in these studies. The five types of features include evolutionary conservation, domain information, network topology, sequence component and expression level. We have described how to implement the useful forms of these features and evaluated their performance based on the data of Escherichia coli MG1655, Bacillus subtilis 168 and human. The prerequisite and applicable range of these features is described. In addition, we have investigated the techniques used to weight features in various models. To facilitate researchers in the field, two available online tools, which are accessible for free and can be directly used to predict gene essentiality in prokaryotes and humans, were referred. This article provides a simple guide for the identification of essential genes in prokaryotes and eukaryotes.

CasLocusAnno: a web-based server for annotating cas loci and their corresponding (sub)types.

In prokaryotes, Clustered Regularly Interspaced Short Palindromic Repeats (CRISPR) and CRISPR-associated protein (Cas) systems constitute adaptive immune systems against mobile genetic elements (MGEs). Here, we introduce the Markov cluster algorithm (MCL) to Makarova et al.'s method in order to select a more reasonable profile. Additionally, our new Maximum Continuous Cas Subcluster (MCCS) method helps identification of tightly clustered loci. The comparison with two other commonly used programs shows that the method could identify Cas proteins with higher accuracy and lower Additional Prediction Rate (APR). Moreover, we developed a web-based server, CasLocusAnno (http://cefg.uestc.cn/CasLocusAnno), capable of annotating Cas proteins, cas loci and their (sub)types less than ~ 28 s following the whole proteome sequence submission. Its standalone version can be downloaded at https://github.com/RiversDong/CasLocusAnno.

Geptop 2.0: An Updated, More Precise, and Faster Geptop Server for Identification of Prokaryotic Essential Genes.

Geptop has performed effectively in the identification of prokaryotic essential genes since its first release in 2013. It estimates gene essentiality for prokaryotes based on orthology and phylogeny. Genome-scale essentiality data of more prokaryotic species are available, and the information has been collected into public essential gene repositories such as DEG and OGEE. A faster and more accurate toolkit is needed to meet the increasing prokaryotic genome data. We updated Geptop by supplementing more validated essentiality data into reference set (from 19 to 37 species), and introducing multi-process technology to accelerate the computing speed. Compared with Geptop 1.0 and other gene essentiality prediction models, Geptop 2.0 can generate more stable predictions and finish the computation in a shorter time. The software is available both as an online server and a downloadable standalone application. We hope that the improved Geptop 2.0 will facilitate researches in gene essentiality and the development of novel antibacterial drugs. The gene essentiality prediction tool is available at http://cefg.uestc.cn/geptop.

Vgas: A Viral Genome Annotation System.

The in-depth study of viral genomes is of great help in many aspects, especially in the treatment of human diseases caused by viral infections. With the rapid accumulation of viral sequencing data, improved, or alternative gene-finding systems have become necessary to process and mine these data. In this article, we present Vgas, a system combining an ab initio method and a similarity-based method to automatically find viral genes and perform gene function annotation. Vgas was compared with existing programs, such as Prodigal, GeneMarkS, and Glimmer. Through testing 5,705 virus genomes downloaded from RefSeq, Vgas demonstrated its superiority with the highest average precision and recall (both indexes were 1% higher or more than the other programs); particularly for small virus genomes (≤ 10 kb), it showed significantly improved performance (precision was 6% higher, and recall was 2% higher). Moreover, Vgas presents an annotation module to provide functional information for predicted genes based on BLASTp alignment. This characteristic may be specifically useful in some cases. When combining Vgas with GeneMarkS and Prodigal, better prediction results could be obtained than with each of the three individual programs, suggesting that collaborative prediction using several different software programs is an alternative for gene prediction. Vgas is freely available at http://cefg.uestc.cn/vgas/ or http://121.48.162.133/vgas/. We hope that Vgas could be an alternative virus gene finder to annotate new genomes or reannotate existing genome.

Identifying Bacterial Essential Genes Based on a Feature-Integrated Method.

Essential genes are those genes of an organism that are considered to be crucial for its survival. Identification of essential genes is therefore of great significance to advance our understanding of the principles of cellular life. We have developed a novel computational method, which can effectively predict bacterial essential genes by extracting and integrating homologous features, protein domain feature, gene intrinsic features, and network topological features. By performing the principal component regression (PCR) analysis for Escherichia coli MG1655, we established a classification model with the average area under curve (AUC) value of 0.992 in ten times 5-fold cross-validation tests. Furthermore, when employing this new model to a distantly related organism-Streptococcus pneumoniae TIGR4, we still got a reliable AUC value of 0.788. These results indicate that our feature-integrated approach could have practical applications in accurately investigating essential genes from broad bacterial species, and also provide helpful guidelines for the minimal cell.

The GC Content as a Main Factor Shaping the Amino Acid Usage During Bacterial Evolution Process.

Understanding how proteins evolve is important, and the order of amino acids being recruited into the genetic codons was found to be an important factor shaping the amino acid composition of proteins. The latest work about the last universal common ancestor (LUCA) makes it possible to determine the potential factors shaping amino acid compositions during evolution. Those LUCA genes/proteins from Methanococcus maripaludis S2, which is one of the possible LUCA, were investigated. The evolutionary rates of these genes positively correlate with GC contents with P-value significantly lower than 0.05 for 94% homologous genes. Linear regression results showed that compositions of amino acids coded by GC-rich codons positively contribute to the evolutionary rates, while these amino acids tend to be gained in GC-rich organisms according to our results. The first principal component correlates with the GC content very well. The ratios of amino acids of the LUCA proteins coded by GC rich codons positively correlate with the GC content of different bacteria genomes, while the ratios of amino acids coded by AT rich codons negatively correlate with the increase of GC content of genomes. Next, we found that the recruitment order does correlate with the amino acid compositions, but gain and loss in codons showed newly recruited amino acids are not significantly increased along with the evolution. Thus, we conclude that GC content is a primary factor shaping amino acid compositions. GC content shapes amino acid composition to trade off the cost of amino acids with bases, which could be caused by the energy efficiency.

Comprehensive exploration of the enzymes catalysing oxygen-involved reactions and COGs relevant to bacterial oxygen utilization.

To better understand the mechanisms of bacterial adaptation in oxygen environments, we explored the aerobic living-associated genes in bacteria by comparing Clusters of Orthologous Groups of proteins' (COGs) frequencies and gene expression analyses and 38 COGs were detected at significantly higher frequencies (p-value less than 1e-6) in aerobes than in anaerobes. Differential expression analyses between two conditions further narrowed the prediction to 27 aerobe-specific COGs. Then, we annotated the enzymes associated with these COGs. Literature review revealed that 14 COGs contained enzymes catalysing oxygen-involved reactions or products involved in aerobic pathways, suggesting their important roles for survival in aerobic environments. Additionally, protein-protein interaction analyses and step length comparisons of metabolic networks suggested that the other 13 COGs may function relevantly with the 14 enzymes-corresponding COGs, indicating that these genes may be highly associated with oxygen utilization. Phylogenetic and evolutionary analyses showed that the 27 COGs did not have similar trees, and all suffered purifying selection pressures. The divergent times of species containing or lacking aerobic COGs validated that the appearing time of oxygen-utilizing gene was approximately 2.80 Gyr ago. In addition to help better understand oxygen adaption, our method may be extended to identify genes relevant to other living environments.