Search

Abstract Background Gene finding is a complicated procedure that encapsulates algorithms for coding sequence modeling, identification of promoter regions, issues concerning overlapping genes and more. In the present study we focus on coding sequence modeling algorithms; that is, algorithms for identification and prediction of the actual coding sequences from genomic DNA. In this respect, we promote a novel multivariate method known as Canonical Powered Partial Least Squares (CPPLS) as an alternative to the commonly used Interpolated Markov model (IMM). Comparisons between the methods were performed on DNA, codon and protein sequences with highly conserved genes taken from several species with different genomic properties. Results The multivariate CPPLS approach classified coding sequence substantially better than the commonly used IMM on the same set of sequences. We also found that the use of CPPLS with codon representation gave significantly better classification results than both IMM with protein (p < 0.001) and with DNA (p < 0.001). Further, although the mean performance was similar, the variation of CPPLS performance on codon representation was significantly smaller than for IMM (p < 0.001). Conclusions The performance of coding sequence modeling can be substantially improved by using an algorithm based on the multivariate CPPLS method applied to codon or DNA frequencies.

Abstract Background We sought to assess whether the concept of relative entropy (information capacity), could aid our understanding of the process of horizontal gene transfer in microbes. We analyzed the differences in information capacity between prokaryotic chromosomes, genomic islands (GI), phages, and plasmids. Relative entropy was estimated using the Kullback-Leibler measure. Results Relative entropy was highest in bacterial chromosomes and had the sequence chromosomes > GI > phage > plasmid. There was an association between relative entropy and AT content in chromosomes, phages, plasmids and GIs with the strongest association being in phages. Relative entropy was also found to be lower in the obligate intracellular Mycobacterium leprae than in the related M. tuberculosis when measured on a shared set of highly conserved genes. Conclusions We argue that relative entropy differences reflect how plasmids, phages and GIs interact with microbial host chromosomes and that all these biological entities are, or have been, subjected to different selective pressures. The rate at which amelioration of horizontally acquired DNA occurs within the chromosome is likely to account for the small differences between chromosomes and stably incorporated GIs compared to the transient or independent replicons such as phages and plasmids.

Abstract Background In genomics, a commonly encountered problem is to extract a subset of variables out of a large set of explanatory variables associated with one or several quantitative or qualitative response variables. An example is to identify associations between codon-usage and phylogeny based definitions of taxonomic groups at different taxonomic levels. Maximum understandability with the smallest number of selected variables, consistency of the selected variables, as well as variation of model performance on test data, are issues to be addressed for such problems. Results We present an algorithm balancing the parsimony and the predictive performance of a model. The algorithm is based on variable selection using reduced-rank Partial Least Squares with a regularized elimination. Allowing a marginal decrease in model performance results in a substantial decrease in the number of selected variables. This significantly improves the understandability of the model. Within the approach we have tested and compared three different criteria commonly used in the Partial Least Square modeling paradigm for variable selection; loading weights, regression coefficients and variable importance on projections. The algorithm is applied to a problem of identifying codon variations discriminating different bacterial taxa, which is of particular interest in classifying metagenomics samples. The results are compared with a classical forward selection algorithm, the much used Lasso algorithm as well as Soft-threshold Partial Least Squares variable selection. Conclusions A regularized elimination algorithm based on Partial Least Squares produces results that increase understandability and consistency and reduces the classification error on test data compared to standard approaches.

Abstract Background Multivariate approaches are important due to their versatility and applications in many fields as it provides decisive advantages over univariate analysis in many ways. Genome wide association studies are rapidly emerging, but approaches in hand pay less attention to multivariate relation between genotype and phenotype. We introduce a methodology based on a BLAST approach for extracting information from genomic sequences and Soft- Thresholding Partial Least Squares (ST-PLS) for mapping genotype-phenotype relations. Results Applying this methodology to an extensive data set for the model yeast Saccharomyces cerevisiae, we found that the relationship between genotype-phenotype involves surprisingly few genes in the sense that an overwhelmingly large fraction of the phenotypic variation can be explained by variation in less than 1% of the full gene reference set containing 5791 genes. These phenotype influencing genes were evolving 20% faster than non-influential genes and were unevenly distributed over cellular functions, with strong enrichments in functions such as cellular respiration and transposition. These genes were also enriched with known paralogs, stop codon variations and copy number variations, suggesting that such molecular adjustments have had a disproportionate influence on Saccharomyces yeasts recent adaptation to environmental changes in its ecological niche. Conclusions BLAST and PLS based multivariate approach derived results that adhere to the known yeast phylogeny and gene ontology and thus verify that the methodology extracts a set of fast evolving genes that capture the phylogeny of the yeast strains. The approach is worth pursuing, and future investigations should be made to improve the computations of genotype signals as well as variable selection procedure within the PLS framework.

Abstract Background Lactobacillus sakei is valuable in the fermentation of meat products and exhibits properties that allow for better preservation of meat and fish. On these substrates, glucose and ribose are the main carbon sources available for growth. We used a whole-genome microarray based on the genome sequence of L. sakei strain 23K to investigate the global transcriptome response of three L. sakei strains when grown on ribose compared with glucose. Results The function of the common regulated genes was mostly related to carbohydrate metabolism and transport. Decreased transcription of genes encoding enzymes involved in glucose metabolism and the L-lactate dehydrogenase was observed, but most of the genes showing differential expression were up-regulated. Especially transcription of genes directly involved in ribose catabolism, the phosphoketolase pathway, and in alternative fates of pyruvate increased. Interestingly, the methylglyoxal synthase gene, which encodes an enzyme unique for L. sakei among lactobacilli, was up-regulated. Ribose catabolism seems closely linked with catabolism of nucleosides. The deoxyribonucleoside synthesis operon transcriptional regulator gene was strongly up-regulated, as well as two gene clusters involved in nucleoside catabolism. One of the clusters included a ribokinase gene. Moreover, hprK encoding the HPr kinase/phosphatase, which plays a major role in the regulation of carbon metabolism and sugar transport, was up-regulated, as were genes encoding the general PTS enzyme I and the mannose-specific enzyme II complex (EIIman). Putative catabolite-responsive element (cre) sites were found in proximity to the promoter of several genes and operons affected by the change of carbon source. This could indicate regulation by a catabolite control protein A (CcpA)-mediated carbon catabolite repression (CCR) mechanism, possibly with the EIIman being indirectly involved. Conclusions Our data shows that the ribose uptake and catabolic machinery in L. sakei is highly regulated at the transcription level. A global regulation mechanism seems to permit a fine tuning of the expression of enzymes that control efficient exploitation of available carbon sources.

Abstract Background Enterococci rank among the leading causes of nosocomial infections. The failure to identify pathogen-specific genes in Enterococcus faecalis has led to a hypothesis where the virulence of different strains may be linked to strain-specific genes, and where the combined endeavor of the different gene-sets result in the ability to cause infection. Population structure studies by multilocus sequence typing have defined distinct clonal complexes (CC) of E. faecalis enriched in hospitalized patients (CC2, CC9, CC28 and CC40). Results In the present study, we have used a comparative genomic approach to investigate gene content in 63 E. faecalis strains, with a special focus on CC2. Statistical analysis using Fisher's exact test revealed 252 significantly enriched genes among CC2-strains. The majority of these genes were located within the previously defined mobile elements phage03 (n = 51), efaB5 (n = 34) and a vanB associated genomic island (n = 55). Moreover, a CC2-enriched genomic islet (EF3217 to -27), encoding a putative phage related element within the V583 genome, was identified. From the draft genomes of CC2-strains HH22 and TX0104, we also identified a CC2-enriched non-V583 locus associated with the E. faecalis pathogenicity island (PAI). Interestingly, surface related structures (including MSCRAMMs, internalin-like and WxL protein-coding genes) implicated in virulence were significantly overrepresented (9.1%; p = 0.036, Fisher's exact test) among the CC2-enriched genes. Conclusion In conclusion, we have identified a set of genes with potential roles in adaptation or persistence in the hospital environment, and that might contribute to the ability of CC2 E. faecalis isolates to cause disease.

Abstract Background Classification of bacteria within the genus Brucella has been difficult due in part to considerable genomic homogeneity between the different species and biovars, in spite of clear differences in phenotypes. Therefore, many different methods have been used to assess Brucella taxonomy. In the current work, we examine 32 sequenced genomes from genus Brucella representing the six classical species, as well as more recently described species, using bioinformatical methods. Comparisons were made at the level of genomic DNA using oligonucleotide based methods (Markov chain based genomic signatures, genomic codon and amino acid frequencies based comparisons) and proteomes (all-against-all BLAST protein comparisons and pan-genomic analyses). Results We found that the oligonucleotide based methods gave different results compared to that of the proteome based methods. Differences were also found between the oligonucleotide based methods used. Whilst the Markov chain based genomic signatures grouped the different species in genus Brucella according to host preference, the codon and amino acid frequencies based methods reflected small differences between the Brucella species. Only minor differences could be detected between all genera included in this study using the codon and amino acid frequencies based methods. Proteome comparisons were found to be in strong accordance with current Brucella taxonomy indicating a remarkable association between gene gain or loss on one hand and mutations in marker genes on the other. The proteome based methods found greater similarity between Brucella species and Ochrobactrum species than between species within genus Agrobacterium compared to each other. In other words, proteome comparisons of species within genus Agrobacterium were found to be more diverse than proteome comparisons between species in genus Brucella and genus Ochrobactrum. Pan-genomic analyses indicated that uptake of DNA from outside genus Brucella appears to be limited. Conclusions While both the proteome based methods and the Markov chain based genomic signatures were able to reflect environmental diversity between the different species and strains of genus Brucella, the genomic codon and amino acid frequencies based comparisons were not found adequate for such comparisons. The proteome comparison based phylogenies of the species in genus Brucella showed a surprising consistency with current Brucella taxonomy.

Abstract Background Bacterial genomes possess varying GC content (total guanines (Gs) and cytosines (Cs) per total of the four bases within the genome) but within a given genome, GC content can vary locally along the chromosome, with some regions significantly more or less GC rich than on average. We have examined how the GC content varies within microbial genomes to assess whether this property can be associated with certain biological functions related to the organism's environment and phylogeny. We utilize a new quantity GCVAR, the intra-genomic GC content variability with respect to the average GC content of the total genome. A low GCVAR indicates intra-genomic GC homogeneity and high GCVAR heterogeneity. Results The regression analyses indicated that GCVAR was significantly associated with domain (i.e. archaea or bacteria), phylum, and oxygen requirement. GCVAR was significantly higher among anaerobes than both aerobic and facultative microbes. Although an association has previously been found between mean genomic GC content and oxygen requirement, our analysis suggests that no such association exits when phylogenetic bias is accounted for. A significant association between GCVAR and mean GC content was also found but appears to be non-linear and varies greatly among phyla. Conclusions Our findings show that GCVAR is linked with oxygen requirement, while mean genomic GC content is not. We therefore suggest that GCVAR should be used as a complement to mean GC content.

Abstract Background The size of the core- and pan-genome of bacterial species is a topic of increasing interest due to the growing number of sequenced prokaryote genomes, many from the same species. Attempts to estimate these quantities have been made, using regression methods or mixture models. We extend the latter approach by using statistical ideas developed for capture-recapture problems in ecology and epidemiology. Results We estimate core- and pan-genome sizes for 16 different bacterial species. The results reveal a complex dependency structure for most species, manifested as heterogeneous detection probabilities. Estimated pan-genome sizes range from small (around 2600 gene families) in Buchnera aphidicola to large (around 43000 gene families) in Escherichia coli. Results for Echerichia coli show that as more data become available, a larger diversity is estimated, indicating an extensive pool of rarely occurring genes in the population. Conclusion Analyzing pan-genomics data with binomial mixture models is a way to handle dependencies between genomes, which we find is always present. A bottleneck in the estimation procedure is the annotation of rarely occurring genes.

Abstract Background Enterococcus faecalis, traditionally considered a harmless commensal of the intestinal tract, is now ranked among the leading causes of nosocomial infections. In an attempt to gain insight into the genetic make-up of commensal E. faecalis, we have studied genomic variation in a collection of community-derived E. faecalis isolated from the feces of Norwegian infants. Results The E. faecalis isolates were first sequence typed by multilocus sequence typing (MLST) and characterized with respect to antibiotic resistance and properties associated with virulence. A subset of the isolates was compared to the vancomycin resistant strain E. faecalis V583 (V583) by whole genome microarray comparison (comparative genomic hybridization (CGH)). Several of the putative enterococcal virulence factors were found to be highly prevalent among the commensal baby isolates. The genomic variation as observed by CGH was less between isolates displaying the same MLST sequence type than between isolates belonging to different evolutionary lineages. Conclusion The variations in gene content observed among the investigated commensal E. faecalis is comparable to the genetic variation previously reported among strains of various origins thought to be representative of the major E. faecalis lineages. Previous MLST analysis of E. faecalis have identified so-called high-risk enterococcal clonal complexes (HiRECC), defined as genetically distinct subpopulations, epidemiologically associated with enterococcal infections. The observed correlation between CGH and MLST presented here, may offer a method for the identification of lineage-specific genes, and may therefore add clues on how to distinguish pathogenic from commensal E. faecalis. In this work, information on the core genome of E. faecalis is also substantially extended.

Abstract Background Existing methods for analyzing bacterial CGH data from two-color arrays are based on log-ratios only, a paradigm inherited from expression studies. We propose an alternative approach, where microarray signals are used in a different way and sequence identity is predicted using a supervised learning approach. Results A data set containing 32 hybridizations of sequenced versus sequenced genomes have been used to test and compare methods. A ROC-analysis has been performed to illustrate the ability to rank probes with respect to Present/Absent calls. Classification into Present and Absent is compared with that of a gaussian mixture model. Conclusion The results indicate our proposed method is an improvement of existing methods with respect to ranking and classification of probes, especially for multi-genome arrays.

Abstract Background Array-based comparative genome hybridization (aCGH) is a tool for rapid comparison of genomes from different bacterial strains. The purpose of such analysis is to detect highly divergent or absent genes in a sample strain compared to an index strain. Development of methods for analyzing aCGH data has primarily focused on copy number abberations in cancer research. In microbial aCGH analyses, genes are typically ranked by log-ratios, and classification into divergent or present is done by choosing a cutoff log-ratio, either manually or by statistics calculated from the log-ratio distribution. As experimental settings vary considerably, it is not possible to develop a classical discriminant or statistical learning approach. Methods We introduce a more efficient method for analyzing microbial aCGH data using a finite mixture model and a data rotation scheme. Using the average posterior probabilities from the model fitted to log-ratios before and after rotation, we get a score for each gene, and demonstrate its advantages for ranking and detecting divergent genes with enlarged specificity and sensitivity. Results The procedure is tested and compared to other approaches on simulated data sets, as well as on four experimental validation data sets for aCGH analysis on fully sequenced strains of Staphylococcus aureus and Streptococcus pneumoniae. Conclusion When tested on simulated data as well as on four different experimental validation data sets from experiments with only fully sequenced strains, our procedure out-competes the standard procedures of using a simple log-ratio cutoff for classification into present and divergent genes.

Age‐related gut bacterial changes during infancy have been widely studied, but it remains still unknown how these changes are associated with immune cell composition. This study's aim was to explore if the temporal development of gut bacteria during infancy prospectively affects immune cell composition. Faecal bacteria and short‐chain fatty acids were analysed from 67 PreventADALL study participants at four timepoints (birth to 12 months) using reduced metagenome sequencing and gas chromatography. Immune cell frequencies were assessed using mass cytometry in whole blood samples at 12 months. The infants clustered into four groups based on immune cell composition: clusters 1 and 2 showed a high relative abundance of naïve cells, cluster 3 exhibited increased abundance of classical‐ and non‐classical monocytes and clusters 3 and 4 had elevated neutrophil levels. At all age groups, we did observe significant associations between the gut microbiota and immune cell clusters; however, these were generally from low abundant species. Only at 6 months of age we observed significant associations between abundant (>8%) species and immune cell clusters. Bifidobacterium adolescentis and Porphyromonadaceae are associated with cluster 1, while Bacteroides fragilis and Bifidobacterium longum are associated with clusters 3 and 4 respectively. These species have been linked to T‐cell polarization and maturation. No significant correlations were found between short‐chain fatty acids and immune cell composition. Our findings suggest that abundant gut bacteria at 6 months may influence immune cell frequencies at 12 months, highlighting the potential role of gut microbiota in shaping later immune cell composition. Our manuscript's main contribution to the field is the novel finding of associations between gut bacteria at different age categories to immune cell composition at 1 year, which supports the notion that gut microbiota composition affects immune cell development during the first year of life. [ABSTRACT FROM AUTHOR]

The Abi protein family consists of putative membrane-bound metalloproteases. While they are involved in membrane anchoring of proteins in eukaryotes, little is known about their function in prokaryotes. In some known bacteriocin loci, Abi genes have been found downstream of bacteriocin structural genes (e.g., pln locus from Lactobacillus plantarum and sag locus from Streptococcus pyogenes), where they probably are involved in self-immunity. By modifying the profile hidden Markov model used to select Abi proteins in the Pfam protein family database, we show that this family is larger than presently recognized. Using bacteriocin-associated Abi genes as a means to search for novel bacteriocins in sequenced genomes, seven new bacteriocin-like loci were identified in Gram-positive bacteria. One such locus, from Lactobacillus sakei 23K, was selected for further experimental study, and it was confirmed that the bacteriocin-like genes (skkAB) exhibited antimicrobial activity when expressed in a heterologous host and that the associated Abi gene (skkI) conferred immunity against the cognate bacteriocin. Similar investigation of the Abi gene plnI and the Abi-like gene plnL from L. plantarum also confirmed their involvement in immunity to their cognate bacteriocins (PlnEF and PlnJK, respectively). Interestingly, the immunity genes from these three systems conferred a high degree of cross-immunity against each other's bacteriocins, suggesting the recognition of a common receptor. Site-directed mutagenesis demonstrated that the conserved motifs constituting the putative proteolytic active site of the Abi proteins are essential for the immunity function of SkkI, and to our knowledge, this represents a new concept in self-immunity. doi: 10.1128/JB.01553-09

A microarray-based comparative genomic hybridization is used to compare the genomic diversity among Enterococcus faecalis strains and Enterococcus faecalis V583. The results have shown that the genes related to stress tolerance is similar in V583 and Enterococcus faecalis strains, suggesting that Enterococcus faecalis has the ability to resist harsh conditions.