New insights into other importantPublisher: NPG; Journal: Nature: Nature; Article Type: Biology letter DOI: 10.1038/nature06269Page 2 of 33 symbiotic functions including H 2 metabolism, CO 2 -reductive acetogenesis and N 2 fixation are also provided by this first system-wide gene analysis of a microbial community specialized towards plant lignocellulose degradation. Our results underscore how complex even a 1-μl environment can be.All known termite species form obligate, nutritional mutualisms with diverse gut microbial species found nowhere else in nature 3 . Despite nearly a century of study, however, science still has only a meagre understanding of the exact roles of the host and symbiotic microbiota in the complex processes of lignocellulose degradation and conversion. Especially conspicuous is our poor understanding of the hindgut communities of wood-feeding 'higher'termites, the most species-rich and abundant of all termite lineages 4 . Higher termites do not contain hindgut flagellate protozoa, which have long been known to be sources of cellulases and hemicellulases in the 'lower' termites. The host tissue of all wood-feeding termites is known to be the source of one cellulase, a single-domain glycohydrolase family 9 enzyme that is secreted and active in the anterior compartments of the gut tract 5 . Only in recent years has research provided support for a role of termite gut bacteria in the production of relevant hydrolytic enzymes. That evidence includes the observed tight attachment of bacteria to wood particles, the antibacterial sensitivity of particle-bound cellulase activity 2 , and the discovery of a gene encoding a novel endoxylanase (glycohydrolase family 11) from bacterial DNA harvested from the gut tract of a Nasutitermes species 6 . Here, in an effort to learn about gene-centred details relevant to the diverse roles of bacterial symbionts in these successful wood-degrading insects,we initiated a metagenomic analysis of a wood-feeding 'higher' termite hindgut community, performed a proteomic analysis with clarified gut fluid from the same sample, and examined a set of candidate enzymes identified during the course of the study for demonstrable cellulase activity.A nest of an arboreal species closely related to Nasutitermes ephratae and N. corniger ( Supplementary Fig. 1) was collected near Guápiles, Costa Rica. From worker specimens, luminal contents were sampled specifically from the largest hindgut compartment, the microbedense, microlitre-sized region alternatively known as the paunch or the third proctodeal segment (P3; Fig. 1a). In the interest of interpretive clarity, we specifically excluded sampling from and analysis of the microbiota attached to the P3 epithelium and the other distinct microbial communities associated with the other hindgut compartments.Publisher: NPG; Journal: Nature: Nature; Article Type: Biology letter DOI: 10.1038/nature06269Page 3 of 33Total community DNA from pooled P3 luminal contents was purified, cloned and sequenced. About 71 million base pairs of Sang...

Viruses are the most abundant biological entities on Earth, but challenges in detecting, isolating, and classifying unknown viruses have prevented exhaustive surveys of the global virome. Here we analysed over 5 Tb of metagenomic sequence data from 3,042 geographically diverse samples to assess the global distribution, phylogenetic diversity, and host specificity of viruses. We discovered over 125,000 partial DNA viral genomes, including the largest phage yet identified, and increased the number of known viral genes by 16-fold. Half of the predicted partial viral genomes were clustered into genetically distinct groups, most of which included genes unrelated to those in known viruses. Using CRISPR spacers and transfer RNA matches to link viral groups to microbial host(s), we doubled the number of microbial phyla known to be infected by viruses, and identified viruses that can infect organisms from different phyla. Analysis of viral distribution across diverse ecosystems revealed strong habitat-type specificity for the vast majority of viruses, but also identified some cosmopolitan groups. Our results highlight an extensive global viral diversity and provide detailed insight into viral habitat distribution and host-virus interactions.

No abstract

We present GenePRIMP (Gene Prediction IMprovement Pipeline, http://geneprimp.jgipsf.org), a computational process that performs evidence-based evaluation of gene models in prokaryotic genomes and reports anomalies including inconsistent start sites, missed genes, and split genes. We show that manual curation of gene models using the anomaly reports generated by GenePRIMP improves their quality and demonstrate the applicability of GenePRIMP in improving finishing quality and comparing different genome sequencing and annotation technologies.More than 1000 microbial genomes have been completely sequenced to date 1 . The increasing number of sequencing projects driven by high-throughput sequencing technologies has further underscored the importance of computational methods in annotating and mining genomic data. For any genome, gene finding is the key step to understanding the biochemistry, physiology, and ecology of the organism. Gene finding relies heavily on computational methods and very few sequencing projects are complemented by the experimental verification of computationally predicted genes through functional genomics experiments or mapping of N-terminal sequences 2,3 . Together with multiple sequencing technologies, multiple gene finders, and somewhat imprecise standards for the identification of genes, this can result in different researchers arriving at substantially varying gene models for the same organism 4 (Fig. 1, Table 1). Consequently, higher standards of accuracy are required for computational gene prediction tools.The most popular gene finders are ab initio and work by statistically profiling protein coding, intergenic, and boundary regions using a variety of classifiers. While most ab initio gene callers boast an average accuracy of 90% or better [5][6][7] , accuracy can be compromised by many factors such as genomic islands of differing GC content, pseudogenes, and genes with programmed or artificial frameshifts, leading to sizeable variability between their gene model predictions. To improve gene models generated by ab initio predictions, some tools include heuristics and post-processing steps such as overlap removal, translation initiation site adjustment, and frameshift detection 8,9 , while others rely on the presence of sequenced close relatives 10 or experimental evidence 11,12 . However, many of these post-processing tools have been tested only on metazoan genomes and use criteria that are not applicable to prokaryotes, and/or are too slow or expensive to perform on a large number of microbial genomes.To overcome the aforesaid limitations of ab initio gene prediction methods, and to address the problem of large variation among their gene models, we have devised GenePRIMP; a computational evidence-based post-processing pipeline that identifies erroneously predicted genes. Manual correction of GenePRIMP-reported genes results in a standardized output gene complement for an organism (sequence) irrespective of the method used for initial gene predictions (Fig.

Fibrobacter succinogenes is an important member of the rumen microbial community that converts plant biomass into nutrients usable by its host. This bacterium, which is also one of only two cultivated species in its phylum, is an efficient and prolific degrader of cellulose. Specifically, it has a particularly high activity against crystalline cellulose that requires close physical contact with this substrate. However, unlike other known cellulolytic microbes, it does not degrade cellulose using a cellulosome or by producing high extracellular titers of cellulase enzymes. To better understand the biology of F. succinogenes, we sequenced the genome of the type strain S85 to completion. A total of 3,085 open reading frames were predicted from its 3.84 Mbp genome. Analysis of sequences predicted to encode for carbohydrate-degrading enzymes revealed an unusually high number of genes that were classified into 49 different families of glycoside hydrolases, carbohydrate binding modules (CBMs), carbohydrate esterases, and polysaccharide lyases. Of the 31 identified cellulases, none contain CBMs in families 1, 2, and 3, typically associated with crystalline cellulose degradation. Polysaccharide hydrolysis and utilization assays showed that F. succinogenes was able to hydrolyze a number of polysaccharides, but could only utilize the hydrolytic products of cellulose. This suggests that F. succinogenes uses its array of hemicellulose-degrading enzymes to remove hemicelluloses to gain access to cellulose. This is reflected in its genome, as F. succinogenes lacks many of the genes necessary to transport and metabolize the hydrolytic products of non-cellulose polysaccharides. The F. succinogenes genome reveals a bacterium that specializes in cellulose as its sole energy source, and provides insight into a novel strategy for cellulose degradation.