Protein Sequence Search based on N-gram Indexing

Percolation of annotation errors through hierarchically structured protein sequence databases

Duplication in biological sequence databases has persisted for 20 years. Duplicate records introduce redundancies to databases, delay biocuration processes, and undermine the accuracy of studies based on sequence analysis such as GC content and melting temperature. Rapid growth of data makes purely manual de-duplication nearly impossible, and existing automatic systems cannot detect duplicates as precisely as experts. Supervised learning has the potential to address such problems by building automatic systems that learn from expert curation to detect duplicates precisely and efficiently. While a mature approach in other duplicate detection contexts, machine learning has seen only preliminary application in the large biological sequence databases. We developed a supervised duplicate detection method, employing an over one million-pair expert curated dataset of duplicates across five organisms derived from genomic sequence databases. We selected 22 features to represent distinct attributes of the database records, and developed both binary and multi-class models. Both models achieve promising performance; the binary model had over 90% accuracy in all the 5 organisms while the multi-class model maintains high accuracy and is more robust in generalisation. We performed an ablation study to quantify the impact of different sequence record features, finding that features derived from meta-data, sequence identity, and alignment quality impact performance most strongly. In particular, better measurement on sequences drives the performance.

BackgroundThe Smith-Waterman algorithm is one of the most widely used tools for searching biological sequence databases due to its high sensitivity. Unfortunately, the Smith-Waterman algorithm is computationally demanding, which is further compounded by the exponential growth of sequence databases. The recent emergence of many-core architectures, and their associated programming interfaces, provides an opportunity to accelerate sequence database searches using commonly available and inexpensive hardware.FindingsOur CUDASW++ implementation (benchmarked on a single-GPU NVIDIA GeForce GTX 280 graphics card and a dual-GPU GeForce GTX 295 graphics card) provides a significant performance improvement compared to other publicly available implementations, such as SWPS3, CBESW, SW-CUDA, and NCBI-BLAST. CUDASW++ supports query sequences of length up to 59K and for query sequences ranging in length from 144 to 5,478 in Swiss-Prot release 56.6, the single-GPU version achieves an average performance of 9.509 GCUPS with a lowest performance of 9.039 GCUPS and a highest performance of 9.660 GCUPS, and the dual-GPU version achieves an average performance of 14.484 GCUPS with a lowest performance of 10.660 GCUPS and a highest performance of 16.087 GCUPS.ConclusionCUDASW++ is publicly available open-source software. It provides a significant performance improvement for Smith-Waterman-based protein sequence database searches by fully exploiting the compute capability of commonly used CUDA-enabled low-cost GPUs.

Identification of new protein coding sequences and signal peptidase cleavage sites of Helicobacter pylori strain 26695 by proteogenomics

Peptide identification by tandem mass spectrometry is the dominant proteomics workflow for protein characterization in complex samples. Traditional search engines, which match peptide sequences with tandem mass spectra to identify the samples' proteins, use protein sequence databases to suggest peptide candidates for consideration. Although the acquisition of tandem mass spectra is not biased toward well-understood protein isoforms, this computational strategy is failing to identify peptides from alternative splicing and coding SNP protein isoforms despite the acquisition of good-quality tandem mass spectra. We propose, instead, that expressed sequence tags (ESTs) be searched. Ordinarily, such a strategy would be computationally infeasible due to the size of EST sequence databases; however, we show that a sophisticated sequence database compression strategy, applied to human ESTs, reduces the sequence database size approximately 35-fold. Once compressed, our EST sequence database is comparable in size to other commonly used protein sequence databases, making routine EST searching feasible. We demonstrate that our EST sequence database enables the discovery of novel peptides in a variety of public data sets.

The role of UniProt's protein sequence databases in biomedical research

Objective To assess the clinical application values of the protein markers associated with Down syndrome （DS）in maternal serum which were screened and identified. Methods Seven maternal serum samples with DS fetus （DS group） and 7 maternal serum samples with normal fetus （control group） in the second trimester were separated by two-dimensional gel electrophoresis （2-DE）.The differentinal expression profile of proteome in maternal serum from DS group was established.The differentially expressed proteins were screened by mass spectrometry （MS） and some proteins were verified by Western blotting （WB）. Results Twenty-nine proteins were discovered to be differentially expressed by more than 1.5 folds in maternal serum from DS group, among which 19 proteins were up-regulated and 10 proteins were down- regulated.Eight proteins displayed 2 or more folds changes in maternal serum from DS group were identified by MS and possibly matched with 12 proteins in Ameracan National Center of Biotchnology Information （NCBI） protein sequence database, such as dGTPase and Beta2-Glycoprotein I （β2-GPI）, etc.The results of WB showed that the mean a values of dGTPase and β2-GPI were 21 567.0±3009.4 and 22 097.0±3958.9 in the DS group, 3957.7±250.9 and 1799.7±105.5 in the control group respectively, which presented that the expression of dGTPase and β2-GPI significantly increased in DS group (t′dGTPase=-17.66，t′β2-GPI=-14.83，P＜0.0001). Conclusions 2-DE and MS are effective methods for preliminary identification of protein markers associated with DS in maternal serum.dGTPase and β2-GPI verified by WB laid a solid fundation for further screening new biologic markers for early diaglosis of DS.(Chin J Lab Med， 2012,35：328-332) Key words: Down syndrome; Blood proteins; Biological markers; Proteomics

Adrenodoxin reductase is an NADP dependent flavoenzyme which functions as the reductase of mitochondrial P 450 systems. We sequenced two adrenodoxin reductase cDNAs isolated from a bovine adrenal cortex cDNA library. The deduced amino acid sequence shows no similarity to the sequence of the microsomal P 450 systems or other known protein sequences. Nonetheless, by sequence analysis and c comparisons with known sequences of dinucleotide-binding folds of two NADP-binding flavoenzymes, two regions of adrenodoxin reductase sequence were identified as the FAD- and NADP-binding sites. These analyses revealed a consensus sequence for the NADP-binding dinucleotide fold (GXGXXAXXXAXXXXXXG, in one-letter amino acid code) that differs from FAD and NAD-binding dinucleotide-fold sequences. In the data base of protein sequences, the NADP-binding-site sequence appears solely in NADP-dependent enzymes, the binding sites of which were not known to date. Thus, this sequence may be used for identification of a certain type of NADP-binding site of enzymes that show no significant sequence similarity.

2 - Nucleic Acid and Protein Sequence Databases

SBASE 5.0 is the fifth release of SBASE, a collection of annotated protein domain sequences that represent various structural, functional, ligand-binding and topogenic segments of proteins. SBASE was designed to facilitate the detection of functional homologies and can be searched with standard database-search programs. The present release contains over 79863 entries provided with standardized names and is cross-referenced to all major sequence databases and sequence pattern collections. The information is assigned to individual domains rather than to entire protein sequences, thus SBASE contains substantially more cross-references and links than do the protein sequence databases. The entries are clustered into >16 000 groups in order to facilitate the detection of distant similarities. SBASE 5.0 is freely available by anonymous 'ftp' file transfer from <ftp.icgeb.trieste.it >. Automated searching of SBASE with BLAST can be carried out with the WWW-server <http://www.icgeb.trieste.it/sbase/ >. and with the electronic mail server <sbase@icgeb.trieste.it >which now also provides a graphic representation of the homologies. A related WWW-server <http://www.abc.hu/blast.html > and e-mail server <domain@hubi.abc.hu > predicts SBASE domain homologies on the basis of SWISS-PROT searches.

If you have just identified a protein binding to a novel regulatory sequence, the next step is to clone the cDNA encoding the protein. Most practitioners in this field would first attempt direct screening of expression libraries with the DNA binding site (see Chapter 5) before embarking on the protein purification/peptide sequence/oligonucleotide screening approach. However, in many cases, direct screening of libraries has not proved successful or has not been possible (for example when the protein is part of a multiprotein complex and more than one protein is required for DNA binding) and protein purification may then have been the only method available. The purification of DNA binding proteins has been remarkably successful with the development of sequence-specific DNA affinity chromatography and the band-shift technique for rapidly analysing chromatography fractions. With the recent advances in protein-sequencing techniques (e.g. sequencing peptides immobilized on blots following electrophoretic separation), the protein purification approach has become less arduous and one can expect to get good sequence results from 50 pmol of protein. Also, mass spectroscopy techniques are being used to mass fingerprint proteins for the identification of protein sequences in databases, and tandem mass spectroscopy is now able to obtain peptide sequences from femtomole amounts of protein.

Bacteriophage piR1-37 was isolated based on its ability to infect strain YeO3-R1, a virulence-plasmid-cured O antigen-negative derivative of Yersinia enterocolitica serotype O : 3. In this study, the phage receptor was found to be a structure in the outer core hexasaccharide of Y. enterocolitica O : 3 LPS. The phage receptor was present in the outer core of strains of many other Y. enterocolitica serotypes, but also in some Yersinia intermedia strains. Surprisingly, the receptor structure resided in the O antigen of Yersinia pseudotuberculosis O : 9. Electron microscopy demonstrated that phiR1-37 particles have an icosahedral head of 88 nm, a short neck of 10 nm, a long contractile tail of 236 nm, and tail fibres of at least 86 nm. This implies that the phage belongs to the order Caudovirales and the family Myoviridae in the ICTV (International Committee for Taxonomy of Viruses) classification. phiR1-37 was found to have a lytic life cycle, with eclipse and latent periods of 40 and 50 min, respectively, and a burst size of approximately 80 p.f.u. per infected cell. Restriction digestions and PFGE showed that the phiR1-37 genome was dsDNA and approximately 270 kb in size. Enzymically hydrolysed DNA was subjected to HPLC-MS/MS analysis, which demonstrated that the phiR1-37 genome is composed of DNA in which thymidine (T) is >99 % replaced by deoxyuridine (dU). The only organisms known to have similar DNA are the Bacillus subtilis-specific bacteriophages PBS1 and PBS2. N-terminal amino acid sequences of four major structural proteins did not show any similarity to (viral) protein sequences in databases, indicating that close relatives of phiR1-37 have not yet been characterized. Genes for two of the structural proteins, p24 and p46, were identified from the partially sequenced phiR1-37 genome.

Gene Clustering is one of the most popular application in the field of Bioinformatics. It is a method of grouping gene into clusters, such that each cluster must have similar gene expression levels. The two most popular population based globalized search algorithms Particle Swarm Optimization (PSO) and Genetic Algorithm (GA) are used for clustering gene expression data but the main drawback of these two algorithm's is that they got trapped in local optima while exploring across the problem space. So the PSO and GA has been hybridized with several other methods, to overcome the local optima problem. This paper presents a literature survey of both GA and PSO's application in Gene Clustering. PSO and GA variants and their hybridization with other metaheuristics are also described in this paper. An attempt is made to provide a guide for the researchers who are working in the area of population based gene clustering. Bioinformatics field deals with the analysis of the biological data, which stored information in the form of DNA, protein sequence in various biological databases. The Bioinformatics field involves various problem such as Gene Clustering Problem, Molecular Docking Problem, Multiple Sequence Alignment Problems, Phylogenetic tree construction , RNA Secondary Structure Prediction, Protein Secondary Structure Prediction, Fragment Assembly Problem. In Gene Clustering the gene is defined as a sequence of nucleotide bases, which carry the required information for the purpose of protein formation, which helps in the construction of the structural components. Clustering is a process of combining input patterns into groups based on the similar functionality. So, Gene clustering is a process of grouping genes in the same clusters based on similar gene expression levels. Gene expression refers to a process through which the coded information from a gene is converted into structures operating in the cell. In order to cluster the Gene expression data , various population based algorithms were used by a number of researchers. The two most popular algorithms among these are PSO (Particle Swarm Optimization) and

In high-throughput proteomics the development of computational methods and novel experimental strategies often rely on each other. In certain areas, mass spectrometry methods for data acquisition are ahead of computational methods to interpret the resulting tandem mass spectra. Particularly, although there are numerous situations in which a mixture tandem mass spectrum can contain fragment ions from two or more peptides, nearly all database search tools still make the assumption that each tandem mass spectrum comes from one peptide. Common examples include mixture spectra from co-eluting peptides in complex samples, spectra generated from data-independent acquisition methods, and spectra from peptides with complex post-translational modifications. We propose a new database search tool (MixDB) that is able to identify mixture tandem mass spectra from more than one peptide. We show that peptides can be reliably identified with up to 95% accuracy from mixture spectra while considering only a 0.01% of all possible peptide pairs (four orders of magnitude speedup). Comparison with current database search methods indicates that our approach has better or comparable sensitivity and precision at identifying single-peptide spectra while simultaneously being able to identify 38% more peptides from mixture spectra at significantly higher precision.

The MIPS group [Munich Information Center for Protein Sequences of the German National Center for Environment and Health (GSF)] at the Max-Planck-Institute for Biochemistry, Martinsried near Munich, Germany, is involved in a number of data collection activities, including a comprehensive database of the yeast genome, a database reflecting the progress in sequencing the Arabidopsis thaliana genome, the systematic analysis of other small genomes and the collection of protein sequence data within the framework of the PIR-International Protein Sequence Database (described elsewhere in this volume). Through its WWW server (http://www.mips.biochem.mpg.de ) MIPS provides access to a variety of generic databases, including a database of protein families as well as automatically generated data by the systematic application of sequence analysis algorithms. The yeast genome sequence and its related information was also compiled on CD-ROM to provide dynamic interactive access to the 16 chromosomes of the first eukaryotic genome unraveled.