Protein Family Classification Research Articles

PANTHER: a browsable database of gene products organized by biological function, using curated protein family and subfamily classification.

The PANTHER database was designed for high-throughput analysis of protein sequences. One of the key features is a simplified ontology of protein function, which allows browsing of the database by biological functions. Biologist curators have associated the ontology terms with groups of protein sequences rather than individual sequences. Statistical models (Hidden Markov Models, or HMMs) are built from each of these groups. The advantage of this approach is that new sequences can be automatically classified as they become available. To ensure accurate functional classification, HMMs are constructed not only for families, but also for functionally distinct subfamilies. Multiple sequence alignments and phylogenetic trees, including curator-assigned information, are available for each family. The current version of the PANTHER database includes training sequences from all organisms in the GenBank non-redundant protein database, and the HMMs have been used to classify gene products across the entire genomes of human, and Drosophila melanogaster. The ontology terms and protein families and subfamilies, as well as Drosophila gene c;assifications, can be browsed and searched for free. Due to outstanding contractual obligations, access to human gene classifications and to protein family trees and multiple sequence alignments will temporarily require a nominal registration fee. PANTHER is publicly available on the web at http://panther.celera.com.

An efficient algorithm for large-scale detection of protein families

Detection of protein families in large databases is one of the principal research objectives in structural and functional genomics. Protein family classification can significantly contribute to the delineation of functional diversity of homologous proteins, the prediction of function based on domain architecture or the presence of sequence motifs as well as comparative genomics, providing valuable evolutionary insights. We present a novel approach called TRIBE-MCL for rapid and accurate clustering of protein sequences into families. The method relies on the Markov cluster (MCL) algorithm for the assignment of proteins into families based on precomputed sequence similarity information. This novel approach does not suffer from the problems that normally hinder other protein sequence clustering algorithms, such as the presence of multi-domain proteins, promiscuous domains and fragmented proteins. The method has been rigorously tested and validated on a number of very large databases, including SwissProt, InterPro, SCOP and the draft human genome. Our results indicate that the method is ideally suited to the rapid and accurate detection of protein families on a large scale. The method has been used to detect and categorise protein families within the draft human genome and the resulting families have been used to annotate a large proportion of human proteins.

Classification of Transmembrane Protein Families Based on Topological Similarity

Classification of Transmembrane Protein Families Based on Topological Similarity

Structural and evolutionary relationships among protein tyrosine phosphatase domains.

With the current access to the whole genomes of various organisms and the completion of the first draft of the human genome, there is a strong need for a structure-function classification of protein families as an initial step in moving from DNA databases to a comprehensive understanding of human biology. As a result of the explosion in nucleic acid sequence information and the concurrent development of methods for high-throughput functional characterization of gene products, the genomic revolution also promises to provide a new paradigm for drug discovery, enabling the identification of molecular drug targets in a significant number of human diseases. This molecular view of diseases has contributed to the importance of combining primary sequence data with three-dimensional structure and has increased the awareness of computational homology modeling and its potential to elucidate protein function. In particular, when important proteins or novel therapeutic targets are identified—like the family of protein tyrosine phosphatases (PTPs) (reviewed in reference 53)—a structure-function classification of such protein families becomes an invaluable framework for further advances in biomedical science. Here, we present a comparative analysis of the structural relationships among vertebrate PTP domains and provide a comprehensive resource for sequence analysis of phosphotyrosine-specific PTPs.

Efficient data mining of proteins involved in carcinogenesis by functional classification using Interpro

Data mining in most cases relies on and is limited by a biologist's experience and knowledge. It is concentration- and time-intensive. We have developed an original hierarchical protein family classification, based primarily on protein sequence motifs, functional domains and cellular localization. This protein classification schema has been generated by classifying entries in Interpro 1.2, an integrated resource of protein families, domains and sites1. By using this new classification, one can effectively and rapidly mine data on proteins involved in oncogenesis. This classification is especially tailored toward the biopharmaceutical industry and drug discovery efforts, and it has been extensively used in Hyseq's internal data-mining processes. The hierarchical protein classification has 9 main classes, 56 subclasses, and 3,052 Interpro entries. These Interpro entries represent 574 domains, 2,418 families, 46 repeats and 14 post-translational modification sites from clustering PRINTS, PROSITE, ProDom, SWISS-PROT, TrEMBL and Pfam data. We generate Pfam models by multiple protein sequence alignments and express them mathematically in the form of hidden Markov models. To demonstrate the utility of our classification schema, we searched the SWISS-PROT and TrEMBL databases with Pfam models. Of the 31% of protein sequences that had significant Pfam hits, 4,520 sequences were in the cancer subclass. At present 59 Interpro entries exist in this subclass. These entries include breast cancer susceptibility proteins, retinoblastoma protein domains, p53 tumor antigen, Xeroderma pigmentosum proteins and the Burkitt's lymphoma receptor.

MetaFam: a unified classification of protein families. II. Schema and query capabilities.

Protein sequence and family data is accumulating at such a rapid rate that state-of-the-art databases and interface tools are required to aid curators with their classifications. We have designed such a system, MetaFam, to facilitate the comparison and integration of public protein sequence and family data. This paper presents the global schema, integration issues, and query capabilities of MetaFam. MetaFam is an integrated data warehouse of information about protein families and their sequences. This data has been collected into a consistent global schema, and stored in an Oracle relational database. The warehouse implementation allows for quick removal of outdated data sets. In addition to the relational implementation of the primary schema, we have developed several derived tables that enable efficient access from data visualization and exploration tools. Through a series of straightforward SQL queries, we demonstrate the usefulness of this data warehouse for comparing protein family classifications and for functional assignment of new sequences.

MetaFam: a unified classification of protein families. I. Overview and statistics.

Protein sequence classification is becoming an increasingly important means of organizing the voluminous data produced by large-scale genome sequencing projects. At present, there are several independent classification methods. To aid the general classification effort, we have created a unified protein family resource, MetaFam. MetaFam is a protein family classification built upon 10 publicly-accessible protein family databases (Blocks + DOMO, Pfam, PIR-ALN, PRINTS, PROSITE, ProDom, PROTOMAP, SBASE, and SYSTERS). MetaFam's family 'supersets', as we call them, are created automatically using set-theory to compare families among the databases. Families of one database are matched to those in another when the intersection of their members exceeds all other possible family pairings between the two databases. Pairwise family matches are drawn together transitively to create a new list of protein family supersets. MetaFam family supersets have several useful features: (1) each superset contains more members than the families from which it is composed, because each of the component family databases only works with a subset of our full non-redundant set of proteins; (2) conflicting assignments can be pinpointed quickly, since our analysis identifies individual members that are in conflict with the majority consensus; (3) family descriptions that are absent from automated databases can frequently be assigned; (4) statistics have been computed comparing domain boundaries, family size distributions, and overall quality of MetaFam supersets; (5) the supersets have been loaded into a relational database to allow for complex queries and visualization of the connections among families in a superset and the consensus of individual domain members; and (6) the quality of individual supersets has been assessed using numerous quantitative measures such as family consistency, connectedness, and size. We anticipate this new resource will be particularly useful to genomic database curators.

Classification of Eukaryotic 7-tms Transmembrane Proteins by Binary Topology Pattern

G protein-coupled receptors (GPCRs) play extremely important functions in our body which are mainly to transduce chemical signals across cell membranes, therefore GPCRs are one of the target proteins in the biomedical field. GPCRs are found in large numbers in most eukaryotic genomes, although the only completely known three-dimensional structure in these families is that of rhodopsin from bovine at 2.8 angstroms resolution by X-ray diffraction [3]. Most GPCRs share a conserved secondary structure that consists of seven transmembrane (TM) helices with extracellular N-terminal loop region. In this study, we applied our method for functional identification of TM proteins, by using a TM topology pattern of the number of TM segments (tms) and the length of each loop domains [4], to a dataset of eukaryotic 7-tms TM proteins with non-GPCR classes and sub-families classified according to GPCRDB [2] protein family classification.

Functional classification of cNMP-binding proteins and nucleotide cyclases with implications for novel regulatory pathways in Mycobacterium tuberculosis.

We have analyzed the cyclic nucleotide (cNMP)-binding protein and nucleotide cyclase superfamilies using Bayesian computational methods of protein family identification and classification. In addition to the known cNMP-binding proteins (cNMP-dependent kinases, cNMP-gated channels, cAMP-guanine nucleotide exchange factors, and bacterial cAMP-dependent transcription factors), new functional groups of cNMP-binding proteins were identified, including putative ABC-transporter subunits, translocases, and esterases. Classification of the nucleotide cyclases revealed subtle differences in sequence conservation of the active site that distinguish the five classes of cyclases: the multicellular eukaryotic adenylyl cyclases, the eukaryotic receptor-type guanylyl cyclases, the eukaryotic soluble guanylyl cyclases, the unicellular eukaryotic and prokaryotic adenylyl cyclases, and the putative prokaryotic guanylyl cyclases. Phylogenetic distribution of the cNMP-binding proteins and cyclases was analyzed, with particular attention to the 22 complete archaeal and eubacterial genome sequences. Mycobacterium tuberculosis H37Rv and Synechocystis PCC6803 were each found to encode several more putative cNMP-binding proteins than other prokaryotes; many of these proteins are of unknown function. M. tuberculosis also encodes several more putative nucleotide cyclases than other prokaryotic species.

Classification and function prediction of proteins using diagnostic amino acid patterns.

Protein sequence similarities offer a convenient means for the classification and identification of protein families and superfamilies. Frequently, proteins descended from a common ancestor preserve their basic three-dimensional conformations even when they have accumulated large numbers of amino acid substitutions and short insertions or deletions. These may prohibit establishment of homology or evolutionary relationships by traditional global sequence alignment means. Limited regions of sequence similarity can also be the result of evolutionary convergence driven by a need for a common function. Regardless of whether divergent or convergent evolution played a role in the appearance of local sequence similarities, these confined regions of similarity can provide insight into structural and functional relationships of proteins that otherwise fail to show significant similarity by global alignment methods.

Protein Family Classification Based on Searching a Database of Blocks

The most highly conserved regions of proteins can be represented as "blocks" of locally aligned sequence segments. Previously, an automated system was introduced to generate a database of blocks that is searched for local similarities using a sequence query. Here, we describe a method for searching this database that can also reveal significant global similarities. Local and global alignments are scored independently, so they can be used in concert to infer homology. A set of 7082 diverse sequences not represented in the database provided queries for testing this approach. The resulting distributions of scores led to guidelines for interpretation of search data and to the classification of 289 uncatalogued sequences into known groups. Thirty-eight of these relationships appear to be new discoveries. We also show how searching a database of blocks can be used to detect repeated domains and to find distinct cross-family relationships that were missed in searches of sequence databases.

A unique method to represent proteins

A new method, based on the frequency distribution of amino-acid pairs, is developed to represent proteins. The method can accommodate and convey all of the chemical and many of the physical properties of proteins. It also lends itself to classification of organisms from which a protein is obtained (Erhan, 1978) and also to classification of protein families (Erhan, in preparation).