The UniProt-GO Annotation database in 2011

Genome reduction promotes increase in protein functional complexity in bacteria

Obligate pathogenic and endosymbiotic bacteria typically experience gene loss due to functional redundancy, asexuality, and genetic drift. We hypothesize that reduced genomes increase their functional complexity through protein multitasking, in which many genes adopt new roles to counteract gene loss. Comparisons of interaction networks among six bacteria that have varied genome sizes (Mycoplasma pneumoniae, Treponema pallidum, Helicobacter pylori, Campylobacter jejuni, Synechocystis sp., and Mycobacterium tuberculosis) reveal that proteins in small genomes interact with proteins from a wider range of functions than do their orthologs in larger genomes. This suggests that surviving proteins form increasingly complex functional relationships to compensate for genes that are lost.

LIN28A is a suppressor of ER-associated translation in embryonic stem cells

LIN28 plays a critical role in developmental transition, glucose metabolism, and tumorigenesis. At the molecular level, LIN28 is known to repress maturation of let-7 microRNAs and enhance translation of certain mRNAs. In this study, we obtain a genome-wide view of the molecular function of LIN28A in mouse embryonic stem cells by carrying out RNA crosslinking-immunoprecipitation-sequencing (CLIP-seq) and ribosome footprinting. We find that, in addition to let-7 precursors, LIN28A binds to a large number of spliced mRNAs. LIN28A recognizes AAGNNG, AAGNG, and less frequently UGUG, which are located in the terminal loop of a small hairpin. LIN28A is localized to the periendoplasmic reticulum (ER) area and inhibits translation of mRNAs that are destined for the ER, reducing the synthesis of transmembrane proteins, ER or Golgi lumen proteins, and secretory proteins. Our study suggests a selective regulatory mechanism for ER-associated translation and reveals an unexpected role of LIN28A as a global suppressor of genes in the secretory pathway.

SEED servers: high-performance access to the SEED genomes, annotations, and metabolic models

The remarkable advance in sequencing technology and the rising interest in medical and environmental microbiology, biotechnology, and synthetic biology resulted in a deluge of published microbial genomes. Yet, genome annotation, comparison, and modeling remain a major bottleneck to the translation of sequence information into biological knowledge, hence computational analysis tools are continuously being developed for rapid genome annotation and interpretation. Among the earliest, most comprehensive resources for prokaryotic genome analysis, the SEED project, initiated in 2003 as an integration of genomic data and analysis tools, now contains >5,000 complete genomes, a constantly updated set of curated annotations embodied in a large and growing collection of encoded subsystems, a derived set of protein families, and hundreds of genome-scale metabolic models. Until recently, however, maintaining current copies of the SEED code and data at remote locations has been a pressing issue. To allow high-performance remote access to the SEED database, we developed the SEED Servers (http://www.theseed.org/servers): four network-based servers intended to expose the data in the underlying relational database, support basic annotation services, offer programmatic access to the capabilities of the RAST annotation server, and provide access to a growing collection of metabolic models that support flux balance analysis. The SEED servers offer open access to regularly updated data, the ability to annotate prokaryotic genomes, the ability to create metabolic reconstructions and detailed models of metabolism, and access to hundreds of existing metabolic models. This work offers and supports a framework upon which other groups can build independent research efforts. Large integrations of genomic data represent one of the major intellectual resources driving research in biology, and programmatic access to the SEED data will provide significant utility to a broad collection of potential users.

BioLiP: a semi-manually curated database for biologically relevant ligand-protein interactions

BioLiP (http://zhanglab.ccmb.med.umich.edu/BioLiP/) is a semi-manually curated database for biologically relevant ligand–protein interactions. Establishing interactions between protein and biologically relevant ligands is an important step toward understanding the protein functions. Most ligand-binding sites prediction methods use the protein structures from the Protein Data Bank (PDB) as templates. However, not all ligands present in the PDB are biologically relevant, as small molecules are often used as additives for solving the protein structures. To facilitate template-based ligand–protein docking, virtual ligand screening and protein function annotations, we develop a hierarchical procedure for assessing the biological relevance of ligands present in the PDB structures, which involves a four-step biological feature filtering followed by careful manual verifications. This procedure is used for BioLiP construction. Each entry in BioLiP contains annotations on: ligand-binding residues, ligand-binding affinity, catalytic sites, Enzyme Commission numbers, Gene Ontology terms and cross-links to the other databases. In addition, to facilitate the use of BioLiP for function annotation of uncharacterized proteins, a new consensus-based algorithm COACH is developed to predict ligand-binding sites from protein sequence or using 3D structure. The BioLiP database is updated weekly and the current release contains 204 223 entries.

Sequencing and analysis of full-length cDNAs, 5'-ESTs and 3'-ESTs from a cartilaginous fish, the elephant shark (Callorhinchus milii)

Cartilaginous fishes are the most ancient group of living jawed vertebrates (gnathostomes) and are, therefore, an important reference group for understanding the evolution of vertebrates. The elephant shark (Callorhinchus milii), a holocephalan cartilaginous fish, has been identified as a model cartilaginous fish genome because of its compact genome (∼910 Mb) and a genome project has been initiated to obtain its whole genome sequence. In this study, we have generated and sequenced full-length enriched cDNA libraries of the elephant shark using the 'oligo-capping' method and Sanger sequencing. A total of 6,778 full-length protein-coding cDNA and 10,701 full-length noncoding cDNA were sequenced from six tissues (gills, intestine, kidney, liver, spleen, and testis) of the elephant shark. Analysis of their polyadenylation signals showed that polyadenylation usage in elephant shark is similar to that in mammals. Furthermore, both coding and noncoding transcripts of the elephant shark use the same proportion of canonical polyadenylation sites. Besides BLASTX searches, protein-coding transcripts were annotated by Gene Ontology, InterPro domain, and KEGG pathway analyses. By comparing elephant shark genes to bony vertebrate genes, we identified several ancient genes present in elephant shark but differentially lost in tetrapods or teleosts. Only ∼6% of elephant shark noncoding cDNA showed similarity to known noncoding RNAs (ncRNAs). The rest are either highly divergent ncRNAs or novel ncRNAs. In addition to full-length transcripts, 30,375 5'-ESTs and 41,317 3'-ESTs were sequenced and annotated. The clones and transcripts generated in this study are valuable resources for annotating transcription start sites, exon-intron boundaries, and UTRs of genes in the elephant shark genome, and for the functional characterization of protein sequences. These resources will also be useful for annotating genes in other cartilaginous fishes whose genomes have been targeted for whole genome sequencing.

The Mammalian Phenotype Ontology as a unifying standard for experimental and high-throughput phenotyping data

The Mammalian Phenotype Ontology (MP) is a structured vocabulary for describing mammalian phenotypes and serves as a critical tool for efficient annotation and comprehensive retrieval of phenotype data. Importantly, the ontology contains broad and specific terms, facilitating annotation of data from initial observations or screens and detailed data from subsequent experimental research. Using the ontology structure, data are retrieved inclusively, i.e., data annotated to chosen terms and to terms subordinate in the hierarchy. Thus, searching for "abnormal craniofacial morphology" also returns annotations to "megacephaly" and "microcephaly," more specific terms in the hierarchy path. The development and refinement of the MP is ongoing, with new terms and modifications to its organization undergoing continuous assessment as users and expert reviewers propose expansions and revisions. A wealth of phenotype data on mouse mutations and variants annotated to the MP already exists in the Mouse Genome Informatics database. These data, along with data curated to the MP by many mouse mutagenesis programs and mouse repositories, provide a platform for comparative analyses and correlative discoveries. The MP provides a standard underpinning to mouse phenotype descriptions for existing and future experimental and large-scale phenotyping projects. In this review we describe the MP as it presently exists, its application to phenotype annotations, the relationship of the MP to other ontologies, and the integration of the MP within large-scale phenotyping projects. Finally we discuss future application of the MP in providing standard descriptors of the phenotype pipeline test results from the International Mouse Phenotype Consortium projects.

Endosomal crosstalk: meeting points for signaling pathways

Endocytosis participates in downregulating incoming signals, but 'signaling endosomes' may also serve as physical platforms for crosstalk between signaling pathways. Here, we briefly review the role of endosomes in signaling crosstalk and suggest that endosome-associated scaffold proteins mediate this crosstalk. In addition, using a proteome-wide in silico approach - in which we analyze endosome-binding properties and the capacity of candidates to recruit signaling proteins from more than one distinct pathway - we extend the list of putative crosstalk-mediating endosomal scaffolds. Because endosomal crosstalk may be an important systems-level regulator of pathway communication, scaffold proteins that mediate this crosstalk could be potential targets for pharmacological intervention and synthetic engineering.

The regulatory factor PcRFX1 controls the expression of the three genes of β-lactam biosynthesis in Penicillium chrysogenum

Penicillin biosynthesis is subjected to a complex regulatory network of signalling molecules that may serve as model for other secondary metabolites. The information provided by the new "omics" era about Penicillium chrysogenum and the advances in the knowledge of molecular mechanisms responsible for improved productivity, make this fungus an excellent model to decipher the mechanisms controlling the penicillin biosynthetic pathway. In this work, we have characterized a novel transcription factor PcRFX1, which is an ortholog of the Acremonium chrysogenum CPCR1 and Penicillium marneffei RfxA regulatory proteins. PcRFX1 DNA binding sequences were found in the promoter region of the pcbAB, pcbC and penDE genes. We show in this article that these motifs control the expression of the β-galactosidase lacZ reporter gene, indicating that they may direct the PcRFX1-mediated regulation of the penicillin biosynthetic genes. By means of Pcrfx1 gene knock-down and overexpression techniques we confirmed that PcRFX1 controls penicillin biosynthesis through the regulation of the pcbAB, pcbC and penDE transcription. Morphology and development seemed not to be controlled by this transcription factor under the conditions studied and only sporulation was slightly reduced after the silencing of the Pcrfx1 gene. A genome-wide analysis of processes putatively regulated by this transcription factor was carried out in P. chrysogenum. Results suggested that PcRFX1, in addition to regulate penicillin biosynthesis, is also involved in the control of several pathways of primary metabolism.

Current challenges in genome annotation through structural biology and bioinformatics

With the huge volume in genomic sequences being generated from high-throughout sequencing projects the requirement for providing accurate and detailed annotations of gene products has never been greater. It is proving to be a huge challenge for computational biologists to use as much information as possible from experimental data to provide annotations for genome data of unknown function. A central component to this process is to use experimentally determined structures, which provide a means to detect homology that is not discernable from just the sequence and permit the consequences of genomic variation to be realized at the molecular level. In particular, structures also form the basis of many bioinformatics methods for improving the detailed functional annotations of enzymes in combination with similarities in sequence and chemistry.

CombFunc: predicting protein function using heterogeneous data sources

Only a small fraction of known proteins have been functionally characterized, making protein function prediction essential to propose annotations for uncharacterized proteins. In recent years many function prediction methods have been developed using various sources of biological data from protein sequence and structure to gene expression data. Here we present the CombFunc web server, which makes Gene Ontology (GO)-based protein function predictions. CombFunc incorporates ConFunc, our existing function prediction method, with other approaches for function prediction that use protein sequence, gene expression and protein–protein interaction data. In benchmarking on a set of 1686 proteins CombFunc obtains precision and recall of 0.71 and 0.64 respectively for gene ontology molecular function terms. For biological process GO terms precision of 0.74 and recall of 0.41 is obtained. CombFunc is available at http://www.sbg.bio.ic.ac.uk/combfunc.

The 2012 Nucleic Acids Research Database Issue and the online Molecular Biology Database Collection

The 19th annual Database Issue of Nucleic Acids Research features descriptions of 92 new online databases covering various areas of molecular biology and 100 papers describing recent updates to the databases previously described in NAR and other journals. The highlights of this issue include, among others, a description of neXtProt, a knowledgebase on human proteins; a detailed explanation of the principles behind the NCBI Taxonomy Database; NCBI and EBI papers on the recently launched BioSample databases that store sample information for a variety of database resources; descriptions of the recent developments in the Gene Ontology and UniProt Gene Ontology Annotation projects; updates on Pfam, SMART and InterPro domain databases; update papers on KEGG and TAIR, two universally acclaimed databases that face an uncertain future; and a separate section with 10 wiki-based databases, introduced in an accompanying editorial. The NAR online Molecular Biology Database Collection, available at http://www.oxfordjournals.org/nar/database/a/, has been updated and now lists 1380 databases. Brief machine-readable descriptions of the databases featured in this issue, according to the BioDBcore standards, will be provided at the http://biosharing.org/biodbcore web site. The full content of the Database Issue is freely available online on the Nucleic Acids Research web site (http://nar.oxfordjournals.org/).