- Open Access
PhyloFacts: an online structural phylogenomic encyclopedia for protein functional and structural classification
© Krishnamurthy et al.; licensee BioMed Central Ltd. 2006
- Received: 8 May 2006
- Accepted: 14 September 2006
- Published: 14 September 2006
The Berkeley Phylogenomics Group presents PhyloFacts, a structural phylogenomic encyclopedia containing almost 10,000 'books' for protein families and domains, with pre-calculated structural, functional and evolutionary analyses. PhyloFacts enables biologists to avoid the systematic errors associated with function prediction by homology through the integration of a variety of experimental data and bioinformatics methods in an evolutionary framework. Users can submit sequences for classification to families and functional subfamilies. PhyloFacts is available as a worldwide web resource from http://phylogenomics.berkeley.edu/phylofacts.
- Gene Ontology
- Additional Data File
- Protein Structure Prediction
- Protein Function Prediction
- Evidence Code
Computational methods for protein function prediction have been critical in the post-genome era in the functional annotation of literally millions of novel sequences. The standard protocol for sequence functional annotation - transferring the annotation of a database hit to a sequence 'query' based on predicted homology - has been shown to be prone to systematic error [1–3]. The top hit in a sequence database may have a different function to the query due to neofunctionalization stemming from gene duplication , differences in domain structure [5, 6], mutations at key functional positions, or speciation . Annotation errors have been shown to propagate through databases by the application of homology-based annotation transfer [7–9]. While the exact frequency of annotation error is unknown (one published estimate is 8% or higher ), the importance of detecting and correcting existing errors and preventing future errors is undisputed.
An additional complicating factor in annotation transfer by homology is the complete failure of this approach for an average of 30% of the genes in most genomes sequenced: in some cases no homologs can be detected within a particular significance threshold, for instance, a BLAST  expectation (E) value (that is, the number of hits receiving a given score expected by chance alone in the database searched) of 0.001 or less, while in other cases database hits may be labeled as 'hypothetical' or 'unknown'.
With the huge array of bioinformatics software tools and resources available, it might seem unthinkable that functional annotation accuracy would be so difficult to ensure. Rather like the parable of the blind men and the elephant, each tool used separately provides a partial and imperfect picture; taken as a whole, the probable molecular function of the protein, biological process, cellular component, interacting partners, and other aspects of a protein's function can often come into better focus. For instance, annotation transfer from the top BLAST hit may suggest a protein is a receptor-like protein kinase, while domain structure prediction reveals that no kinase domain is present; the two orthogonal analyses prevent mis-annotation of the unknown protein.
In this paper we present PhyloFacts, an online structural phylogenomics encyclopedia containing almost 10,000 'books' for protein families and domains, designed to improve the accuracy and specificity of protein function prediction . PhyloFacts integrates a wide array of biological data and informatics methods for protein families, organized on the basis of structural similarity and by evolutionary relationships. This enables a biologist to examine a rich array of experimental data and bioinformatics predictions for a protein family, and to quickly and accurately infer the function of a protein in an evolutionary context.
PhyloFacts is motivated by two of the biggest lessons of the post-genome era - the power of integrating data and inference tools from different sources, and improved prediction accuracy using consensus approaches in bioinformatics. For instance, protein structure prediction 'meta-servers' making predictions based on a consensus over results retrieved from several independent servers typically have lower error rates than any one server used separately . In the case of protein structure prediction, we can also take advantage of the fact that members of a large diverse protein family tend to share the same three-dimensional structure even when their primary sequence similarity becomes undetectable. This enables us to use another type of consensus approach involving the application of the same method to several different members of the family to boost prediction accuracy (for example, ).
We employ the same basic principles in this resource, by integrating many different prediction methods and sources of experimental data over an evolutionary tree. In cases where attributes are known to persist over long evolutionary distances (such as protein three-dimensional structure), we can integrate predictions over the entire tree to derive a consensus prediction for the family as a whole. In cases where attributes are more restricted in their distribution in the family (for example, ligand recognition among G-protein coupled receptors), inferences will be more circumspect, potentially restricted to strict orthologs. Evolutionary and structural clustering of proteins enables us to integrate these disparate types of data and inference methods effectively, to identify potential errors in database annotations and provide a platform to improve the accuracy of functional annotation overall.
In addition to new methods developed by us for phylogenomic inference, PhyloFacts includes a number of standard bioinformatics methods available publicly. To motivate the need for protein functional classification integrating diverse methods and data in an evolutionary framework, we examine the major classes of bioinformatics methods in turn, and discuss their different pros and cons. Methods designed for predicting the biological process(es) in which a protein participates (for example, bioinformatics approaches such as Phylogenetic Profiles  and Rosetta Stone , analysis of DNA chip array data, and proteomics experiments such as pull-down experiments, yeast two-hybrid data, and so on) are clearly complementary, and will be included in future releases of the PhyloFacts resource.
Database homolog search tools
Database homolog search tools (for example, BLAST, FASTA , and so on) can be blindingly fast, but do not distinguish between local matches and sequences sharing global similarity; they report a score or E-value measuring the significance of the local match between a query sequence and sequences in the database. This can lead to errors when annotations are transferred in toto based on only local similarity. These pairwise sequence comparison methods of homolog detection have also been shown to have limited effectiveness at recognizing remote homologs (distantly related sequences) .
Iterated homology search methods
Iterated homology search methods such as PSI-BLAST  have been developed in recent years. These methods enable larger numbers of sequences to be annotated functionally, albeit with a potentially higher error rate due to divergence in function from their common ancestor.
Domain-based annotation and protein structure prediction
Domain-based annotation and protein structure prediction libraries of profiles or hidden Markov models (HMMs) for functional or structural domains (PFAM , SMART , or Superfamily ) are particularly helpful when a homolog search fails. There are two primary limitations of this approach to functional annotation. First, these statistical models of protein families and domains are typically designed for sensitivity rather than specificity, and thus afford a fairly coarse level of annotation. For example, the PFAM 7TM_1 HMM recognizes a variety of G-protein coupled receptors, irrespective of their ligand specificity. Second, a protein's function is a composite of all its constituent domains; thus, even in cases where each of a protein's domains can be identified, the actual function of the protein may not be elucidated.
Phylogenomic inference was originally designed to address the problem of annotation transfer from paralogous rather than orthologous genes through the construction and analysis of phylogenetic trees overlaid with experimental data. This approach has been shown to enable the highest accuracy in prediction of protein molecular function [21–23], but inherent technical and computational complexity has limited its use. Several attempts at identification of orthologs (for example, Orthostrapper  and RIO ) and at automating phylogenomic inference of molecular function  have been presented, and may lead to more widespread application of this approach.
Prediction of protein localization
Prediction of protein localization is enabled by resources such as the TMHMM  transmembrane prediction server, the TargetP  cellular component prediction server, and the PHOBIUS  integrated signal peptide and transmembrane prediction server. These provide another perspective on a protein's function, and can suggest participation in biological pathways when other data are lacking. Because these methods can rely on fairly weak and non-specific signals (for example, hydrophobic stretches as indicators of membrane localization), both false positive and false negative predictions are not uncommon .
As of 11 July 2006, the PhyloFacts encyclopedia contains 9,710 'books' for protein superfamilies and structural domains. Each book in the PhyloFacts resource contains heterogeneous data for protein families, including a cluster of homologous proteins, multiple sequence alignment, one or more phylogenetic trees, predicted three-dimensional structures, predicted functional subfamilies, taxonomic distributions, Gene Ontology (GO) annotations , PFAM domains, hyperlinks to key literature and other online resources, and annotations provided by biologist experts. Residues conferring family and subfamily specificity are predicted using alignment/evolutionary analyses; these patterns are plotted on three-dimensional structures. HMMs constructed for each family and subfamily enable classification of novel sequences to different functional classes. Details on each aspect of the resource construction are available in the 'Details on Library Construction and Software Tools' section.
Distribution of various book types in PhyloFacts
No. of books in PhyloFacts
Global homology group
The PhyloFacts phylogenomic resource can be used in several ways: sequences can be submitted for protein structure prediction or functional classification, protein family books can be browsed, and data of various types (multiple sequence alignments (MSAs), phylogenetic trees, HMMs, and so on) can be downloaded from the resource.
PhyloFacts includes books focusing on specific protein families or classes. The largest of these series is the PhyloFacts 'Protein Structure Prediction' library, with 5,328 books, each representing either a structural domain from the Astral database  or protein structures from the Protein Data Bank (PDB ). This series enables biologists to obtain predicted structures for submitted proteins. The books in the Protein Structure Prediction library were created using individual structural domains as seeds, gathering homologs from the NR  database using PSI-BLAST or the UCSC SAM  software tools.
The second major book series in PhyloFacts is the 'Animal Proteome Explorer' library, containing 4,226 protein families in the human genome, expanded to include additional homologs from other organisms. Specialized sections of the Animal Proteome Explorer series are devoted to protein families of particular biomedical relevance: G-Protein Coupled Receptors (65 books), Ion Channels (50 books), and Innate Immunity (52 books). The Animal Proteome Explorer series has been constructed using GHGCluster (see section 'Details on Library Construction and Software Tools'). The GPCR library includes books for protein families based on the classification of the GPCRDB .
The 'Plant Disease Resistance Phylogenomic Explorer' forms the third main series of specialized books in PhyloFacts, devoted to protein families involved in plant disease resistance and host-pathogen interaction (105 books). Families in this series include the canonical plant R (resistance) genes, proteins involved in defense signaling and effector proteins from plant pathogens.
These three main divisions are not strictly distinct, and there are some overlaps. For instance, a book for the Toll Interleukin Receptor (TIR) domain (PhyloFacts book ID: bpg002615) is placed in the Protein Structure Prediction library (due to the presence of a solved structure for this family) as well as in the Innate Immunity and Plant Disease Resistance libraries (since TIR domains are found in both plant and animal proteins involved in eukaryotic innate immunity).
Because our recommended protocol for protein function prediction starts with transfer of annotation from globally alignable orthologs (see section 'Functional annotation using PhyloFacts'), a large number of books in PhyloFacts are designated as type Global Homology, and subjected to rigorous quality control (see section 'Details on Library Construction and Software Tools, Defining Book Type'). Standard protein clustering tools typically ignore the issue of global sequence similarity, so that even resources intending to cluster proteins based on global similarity can occasionally fail (for example, the Celera Panther resource  class Leucine-Rich Transmembrane Proteins [PTHR23154] contains proteins with diverse domain structures; Additional data file 1). By contrast, most web servers for protein functional classification provide primarily domain-level analyses (for example, SMART and PFAM). To supplement these analyses, PhyloFacts also provides books for different types of structural similarities across sequences, including short conserved motifs and structural domains.
Comparison of PhyloFacts with other functional classification resources
Analysis of user-submitted sequences
Classification to full-length protein families
Subfamily level classification
Domain level classification
DNA sequence analysis
Batch-mode sequence inputs allowed
Analysis required for phylogenomic inference
Clusters based on full-length protein families
Phylogenetic trees for full-length protein families
GO data for individual sequences
GO data for clusters
GO evidence codes
EC numbers for individual sequences
EC numbers for each cluster
Analyses required for function inference based on structure
Phylogenetic trees for single domains
Clusters based on domains
Predicted three-dimensional structure for a protein family
Predicted critical residues
PDB structure visualization
Transmembrane domain prediction
Signal peptide prediction
Links to PDB
Additional protein family data
Retrieval of relevant literature for individual families
Extended protein family annotation
Clusters of interacting domain families
Graphic displays of related domain architectures
Fractional coverage of genomes
Number of sequences
In an ideal scenario, annotation transfer between a query and homolog would meet three criteria : first, global homology; second, orthology ; and third, supporting experimental evidence for the functional annotation being transferred. In practice, confirming agreement at all three criteria is not always straightforward. Very few sequences have experimentally solved structures; satisfaction of the first condition is, therefore, typically determined by comparison of their predicted domain structures using, for example, PFAM or Conserved Domain  analysis, or by pairwise alignment analysis. Automated determination of orthology is complicated due to incomplete sequencing, gene duplication and loss, errors in gene structure and other issues; for a review see . Satisfying the last condition is equally difficult due to the paucity of sequences with experimentally determined function; our analysis of GO annotations and evidence codes for over 370,000 sequences in the UniProt database  shows <3% to have experimental evidence supporting a functional annotation. (This statistic is based on the analysis of 372,448 UniProt sequences present in the PhyloFacts resource as of June 2005. Two-thirds of these (248,152) had GO annotations, but only 3% of this smaller set had evidence codes indicating experimental support: IDA (inferred from direct assay), IGI (inferred from genetic interaction), IMP (inferred from mutant phenotype), IPI (inferred from physical interaction), and TAS (traceable author statement).)
Books in the PhyloFacts resource are labeled by the level of structural similarity across members (that is, global homology, domain, and so on), and include phylogenetic trees, inferred subfamilies, and GO annotations and evidence codes to enable a biologist to check for agreement at the three criteria for transferring annotations. In cases where a protein of unknown function is placed in a global homology group with an ortholog having experimentally determined function, annotation transfer can proceed with high confidence. In other cases, the biologist can check for experimentally determined function in paralogous genes (bearing in mind that functions may have diverged), or at domain-based clusters, to obtain clues to the molecular function for different regions of a protein of interest. We attempt to accommodate all of these possibilities; a sequence search against the resource may match books representing global homology groups, structural domains, conserved regions, or even short motifs, all of which are presented to the user (Figure 2).
We note that while domain-based annotation is inherently less precise, PhyloFacts does provide predicted functional subfamilies within domain-based books as well as within books representing global homology groups. While annotation transfer across proteins having different overall folds is prone to systematic error, previous results suggest that subfamily classification of sequences aligned along a single common domain can be consistent with the overall domain structure and molecular function of sequences . Our experiments using SCI-PHY to analyze proteins with different overall domain structures also support the same conclusion (unpublished data, Brown DP, Krishnamurthy N, Sjölander K).
In addition to the value PhyloFacts presents to a human investigator, it also provides a framework for the development of a fully automated functional inference system. A new generation of probabilistic methods for inferring molecular function automatically has arisen in recent years (for example, [26, 49, 50]). For instance, SIFTER uses a Bayesian approach to infer a distribution over possible functions in a phylogenetic tree, taking as input a cluster of sequences, a phylogenetic tree, and GO annotations and evidence codes, all of which PhyloFacts collects and integrates in one resource. SIFTER integration is to be available in our next release.
However, technical issues present barriers to the goal of fully automated function prediction (see  for a review). Sequences in a cluster may have different descriptors based on the species of origin; for example, the Drosophila community is likely to use different names for a gene to that used by the Caenorhabditis elegans community, and both are likely to use different terms to those used by investigators working in mouse genomics. The value of a standardized nomenclature, such as that being developed by GO, is obviously important, but significant work remains in this area. An exhaustive thesaurus of equivalent biological terms would be valuable. The sparse nature of experimentally supported molecular functions provides an additional barrier to automated approaches. We discuss these issues further in the section 'Challenges to phylogenomic inference'.
Clustering together proteins based on predictable global homology enables us to analyze a cluster of homologs as a unit and detect potential errors in annotation; database annotation errors tend to stand out as anomalous against a backdrop of otherwise consistent annotations (unless, of course, annotation errors have percolated through the database).
For instance, the Oryza sativa GenBank protein AAR00644 is labeled as a 'putative LRR receptor-like protein kinase'. The canonical structure of receptor-like kinases (RLKs) consists of an extracellular leucine-rich repeat (LRR) region, a transmembrane domain, and a cytoplasmic kinase domain; AAR00644 contains no kinase domain. On the other hand, AAR00644 does match the canonical structure of closely related receptor-like proteins (RLPs), which are structurally very similar to RLKs, except that they terminate with a short cytoplasmic tail, and do not contain a kinase domain . In the PhyloFacts resource, this protein is classified as a member of the global homology group book 'Plant LRR proteins (putative RLPs)' (PhyloFacts book ID: bpg005632), where PFAM domain analysis of the cluster shows no detectable kinase domains.
For a second example, the GenBank sequence AAF19052 labeled as 'neutral human sphingomyelinase'  appears to be neither human nor a sphingomyelinase. Instead it appears to encode a bacterial isochorismate synthase protein. This sequence is classified to the PhyloFacts book 'Isochorismate synthase-related' (Phylofacts book ID: bpg004927), in which this purportedly 'human' sequence is the only representative eukaryote. (Note that even the translated BLAST search of this sequence against the human genome finds no matches.) In this case, both domain structure analysis and analysis of the taxonomic distribution of the globally homologous members of the family help identify the probable error.
Lastly, G-protein coupled receptor (GPCR) classification is notoriously difficult, with many receptors having no known ligand (termed 'orphan receptors'). One such orphan, a GPCR from river lamprey (UniProt: Q9YHY4), is annotated as 'Putative odorant receptor LOR3', based on its expression in the olfactory epithelium . Standard profile/HMM-based analyses (for example, PFAM, SMART and the NCBI CDD) only match this protein to the PFAM 7TM_1 class, containing dozens of subtypes. BLAST analysis shows other putative odorant receptors from river lamprey (submitted by the same authors) as top hits, followed by trace amine receptors. However, analyses of phylogenetic trees containing this sequence show it (and the other putative odorant receptors detected by BLAST) to be located within subtrees containing trace amine receptors (see PhyloFacts books bpg004950, bpg000525 and bpg000543) and to be quite different from experimentally confirmed odorant receptors (Additional data file 2).
Anomalous annotations such as these are often signs that annotation transfer has gone wrong. In other cases, anomalies may be quite real and provide new insights into the evolution of novel functions in a family. Automated anomaly detection faces the same technical barriers as automated functional annotation, including the need for probabilistic inference of gene function, standardized nomenclatures and exhaustive synonym tables of biological terms. At present, these anomalies - whether true functional differences or database annotation errors - are detected manually. In the future we expect automated function prediction methods will enable anomalous annotations to be flagged for expert examination. Protocols will then need to be established by the biological community to correct any errors and to ensure that sequence databases receive corrected annotations.
Clustering sequences for PhyloFacts books
Sequences for structural domain books were gathered using PSI-BLAST and UCSC SAM Target-2K (T2K) . Sequences retrieved for global homology group books are required to share the same overall domain structure (global alignment). We have two tools for this process: FlowerPower (NK, Brown D, KS, unpublished data) and GHGCluster.
FlowerPower is an iterative homolog detection algorithm like PSI-BLAST that retrieves homologs to a seed sequence (or query) and aligns sequences using profile methods. However, instead of using a single profile to identify and align new sequences, FlowerPower uses subfamily identification and subfamily HMM construction to expand the homology cluster in each iteration. Alignment analysis is used to restrict the cluster to match user-specified criteria (for example, global alignment for protein function prediction using phylogenomic inference, and global-local alignment (global to the seed, local to the database hit) for domain-based clustering). Experimental validation of FlowerPower shows it has greater selectivity than BLAST, PSI-BLAST and the UCSC SAM-T2K methods of homolog detection at discriminating sequences with local similarity from those with global similarity. The FlowerPower server is available online .
The Global Homology Group (GHG) Cluster program enables us to cluster a selected sequence database (for example, a genome) into global homology groups, while also including homologs from a second, generally larger, database.
GHGCluster takes two inputs: a set of sequences Q, containing the sequences to be clustered, and a database D to use for expanding the clusters to include globally alignable homologs from other organisms. A superset of sequences, the expansion database E, is created by merging Q and D. To improve run time, E is partitioned into overlapping bins based on sequence length. A seed sequence (query) is chosen from Q and homologs are gathered from its corresponding bin in E, using PSI-BLAST (E-value < 1e-5; user-specified number of iterations). Each hit is assessed for global homology to the query, based on percent identity (≥20%), and bi-directional alignment coverage, that is the fractional aligned length of both seed and hit (ranging from 60% for sequences <100 residues to 85% for sequences of >500 residues). In some cases, PSI-BLAST returns multiple short aligned regions, none of which is long enough to pass the above requirements. In these cases, the failing hits are realigned to an HMM built from the seed, followed by alignment analysis. The seed and any accepted sequences are defined as a cluster and removed from Q (but not E). A new seed is then chosen from Q and the process is iterated until Q is empty.
This procedure results in a set of clusters, some of which may contain the same sequence(s). At this stage we merge compatible overlapping clusters. We rank all pairs of clusters by the number of sequences they have in common and attempt to merge pairs in order. For each pair, we choose the alignment with the greater number of aligned columns and designate this as cluster A; the other cluster becomes B. We build an HMM from the cluster A alignment and use the COACH algorithm  to align the entire cluster B alignment to this HMM (and therefore to cluster A). If the total fraction of gap characters in the merged alignment (that is, no. of gaps/(no. of sequences × no. of columns)) is less than 20%, and the mean percent identity is greater than 20%, the merge is accepted. Otherwise, if the alignment fails based on gap content alone, we trim columns with >30% gaps from the amino and carboxyl termini and reassess: the merge is accepted if cluster sequences align at least 70% of their residues, on average, within the trimmed 'core' alignment and if the core alignment contains >50% of the columns from the merged alignment and <10% gaps. Note, this procedure is computationally efficient, but can produce clusters that fail to meet the criteria set for global homology group books. In these cases, books are flagged for additional automated analysis and manual inspection in order to maintain quality control.
GHGCluster was used to cluster the human genome in the construction of the PhyloFacts Animal Proteome Explorer, including homologs from NR, using three iterations of PSI-BLAST.
Note on clustering splice and allelic variants
Our atomic unit in PhyloFacts clustering is the protein sequence, independent of its origin. Consequently, splice and allelic variants of a gene are not handled differently to genes from entirely different species during clustering, alignment and tree construction (although they would be interpreted differently during a subsequent phylogenomic analysis using the resource). If the variant retains the same domain structure and is globally alignable to other isoforms, it will be included in the global homology group cluster for those genes, otherwise it may end up in a different book in the resource. It should be noted that the presence of different isoforms for a gene can cause difficulties in phylogenomic inference if their common genome locus is not evident. Future releases of PhyloFacts will display this information and create links between different gene isoforms present in different books.
Multiple sequence alignment
The alignment method is selected based on the type of book. For alignments of global homology group proteins, we normally use the MUSCLE software  to realign sequences obtained in clustering; in some cases, we use the SATCHMO software . Both methods have outstanding performance evaluated on benchmark datasets. Alignments of structural domains are taken directly from the clustering algorithm (PSI-BLAST or T2K). The method used is indicated in the 'Book Details' section at the bottom of each book page, under 'Build method notes'. Multiple sequence alignments for all books (except those constructed to model solved three-dimensional structures) are masked to remove columns with >70% gaps prior to phylogenetic tree construction. Alignments are available for the family as a whole and for each subfamily; these can be downloaded or viewed using the Java-based Jalview software . An annotated alignment, indicating SCI-PHY subfamily membership, is also available for viewing and download. Alignment statistics are provided, including average, minimum and maximum percent identity, fraction of gap characters in the MSA, and other relevant measurements.
Defining book type
To be defined as a 'Global Homology' book, a multiple sequence alignment must meet the following criteria: first, ≤15% gap characters over the multiple sequence alignment; second, ≤30% columns with BLOSUM62  sum-of-pairs scores < 0; third, difference between the longest and the shortest sequence in the alignment <150 amino acids; and fourth, all sequences align over ≥75% of their length. Books of type 'Domain' were required to match a structural domain (as determined by SCOP) or to correspond to a PDB structure. Books labeled as 'Conserved Region' required global-local alignment of sequences to the HMM (generally matching over 70% of the HMM match states). Most books labeled as 'Pending' are those that were produced by the GHGCluster program, but which failed the stringent 'Global Homology Group' alignment quality control tests; the final classification of these books to the different structural types is in progress.
Subfamily identification is provided using the SCI-PHY (Subfamily Classification In Phylogenomics) software . SCI-PHY is an automatic subfamily identification algorithm; given an input MSA, SCI-PHY uses Dirichlet mixture densities  and relative entropy to construct a hierarchical tree, and cuts the tree into subtrees to identify subfamilies using minimum-description-length principles. Extensive studies show SCI-PHY subfamilies correspond closely to both expert-identified subtypes and to conserved clades in phylogenetic trees (unpublished data, Brown DP, Krishnamurthy N, Sjölander K). The SCI-PHY server is available online .
HMM construction for the family and individual subfamilies
The UCSC Sequence Alignment and Modeling (SAM) software is used to construct HMMs and in scoring sequences against HMMs . This software was selected based on its outstanding performance in remote homology detection [17, 64]. Family HMMs are constructed using the UCSC SAM w0.5 software. Subfamily HMMs (SHMMs) are constructed as described in . Validation experiments on over 500 unique SCOP folds comparing subfamily and family HMMs show SHMMs to have high specificity in detecting functionally similar sequences and to improve the range of homolog detection with significant scores (unpublished data, Brown DP, Krishnamurthy N, Sjölander K).
Protein structure and domain prediction
PFAM domains are identified using the consensus sequence for the family as a query using the PFAM gathering threshold as a cutoff. Matches to PDB structures are predicted by BLAST analysis of the family consensus sequence, using an E-value cutoff of 0.001 (that is, protein structure prediction based on inferred homology). Any putative homologous structures were aligned to the family HMM using local-local alignment (SAM parameter - sw 2). Transmembrane domains and signal peptides are predicted using the PHOBIUS server , selected due to its ability to differentiate between signal peptides and transmembrane domains.
Phylogenetic tree construction and visualization
Because many of the protein superfamilies in the PhyloFacts resource span extremes of evolutionary divergence (for example, with pairwise identities <20%), tree topologies produced by different methods can often disagree. For this reason, most of the protein families in the resource contain several phylogenetic trees built using different algorithms, enabling biologists interested in these families to examine the differences and commonalities between the trees.
Neighbor-Joining trees are constructed using the PHYLIP software , using the default parameters for 'protdist' and 'neighbor' (JTT model  and no variation of rates). Maximum Likelihood trees are estimated using the PHYML software , also using the JTT model, four substitution rate categories, and a gamma-distributed model of rates (gamma = 1), and are set to optimize tree topology only. Maximum Parsimony trees are estimated using the PAUP* software , by taking an extended majority rule consensus of the most parsimonious trees obtained via ten repetitions of heuristic tree search. All trees are rooted using the midpoint method. As of 4 July 2006, of the 9,707 books in the library, 8,511 have at least one true phylogenetic tree constructed, and 3,613 have NJ, ML and MP trees. All books have had SCI-PHY subfamily analyses completed, and will eventually include Neighbor-Joining (including bootstrap values), Maximum Likelihood and Maximum Parsimony trees.
Predicted critical residues
Novel sequence classification
Classification to a protein family is enabled by HMM scoring. Since HMM scoring is computationally intensive and the PhyloFacts resource contains almost 10,000 books (each containing a family HMM and potentially dozens of subfamily HMMs), we provide heuristic approaches enabling rapid classification of user-submitted sequences. For computational efficiency, we select books for HMM scoring via BLAST analysis of the submitted query against a dataset of over 2.5 million PhyloFacts training sequences using an E-value cutoff of 10; this significantly reduces the number of HMM scores required without affecting sensitivity. Users can override this 'BLAST pre-screen' protocol using the 'Advanced' settings page. Books retrieved based on either protocol can then be selected for scoring the submitted sequence(s) against subfamily HMMs for additional specificity of functional classification.
In future releases of the PhyloFacts resource we plan to include automated predictions of protein function using the SIFTER software. SIFTER will be used to provide predicted molecular functions and participation in biological processes for SCI-PHY subfamilies as well as for conserved evolutionary clades. Links between proteins, or between books, will be provided, to reflect the many types of relationships (for example, participating in the same pathway or complex, sharing a common domain or a predicted common ancestor, splice variation, and so on). Users will be able to navigate between proteins in the same pathway using the Cytoscape software, in which links to PhyloFacts books will be embedded. Users will be able to retrieve comparative (homology) models for selected proteins from the ModBase resource  through hyperlinks on book pages. Literature will be retrieved automatically for sequences in a book, and natural language processing software will be used to summarize the key points. We will expand the resource for improved coverage of key (animal) model organisms (human, mouse, C. elegans, D. melanogaster), and to keep our protein structure prediction library current. We plan to reduce redundancy in the library by combining books with significant sequence overlap. We will develop software tools to identify and include new family members as well as new experimental data (for example, the availability of solved structures, results from site-directed mutagenesis, protein-protein interactions, and so on). We will also include extensions to the subfamily HMM scoring protocol to differentiate between sequences representing a novel subtype and those that can be classified to the top-scoring subfamily (based on logistic regression analysis). Phylogenetic trees for each family will be extended to include strict consensus trees across two or more methods, and bootstrap analysis will be provided for Neighbor Joining trees. Finally, we plan to provide community annotation tools to enable biologists to upload their data, commentaries and hyperlinks to experimental data for members of protein families.
Phylogenomic analysis of protein function is known to improve the accuracy of functional annotation, but has had restricted application due to its technical complexity. The PhyloFacts resource enables biologists who may have limited bioinformatics expertise to take advantage of pre-computed phylogenomic analyses for hundreds of thousands of proteins. New sequences can be classified to families and subfamilies using over 700,000 hidden Markov models, for increased functional specificity. The resource as a whole brings together many different types of bioinformatics analyses and data; the integration of these data and improved orthology determination enables biologists to avoid making new annotation errors at the outset, and to detect, and possibly correct, existing annotation errors.
Phylogenetic uncertainty and ambiguity remain significant challenges to phylogenomic inference of molecular function. Consensus analysis can be used to detect clades with support across two or more tree methods, but this approach could inadvertently be misleading if inherent biases in the methods are not taken into account. It is critical that the computational biology community and the systematics community work together to develop methods to assess the expected accuracy of phylogenetic methods for protein superfamily reconstruction; new simulation studies should be developed that model the kinds of structural and functional changes observed in protein superfamily evolution.
Other challenges to automating functional annotation through phylogenomics methods include: the lack of a standardized nomenclature for gene names across different model organisms, although the GO consortium efforts are making progress in this respect; natural structural and sequence divergence among family members causing difficulties in clustering; ensuring multiple sequence alignment accuracy when divergently related sequences are included in a cluster; and the persistence of database annotation errors. Database annotation errors should be correctable at the source, so that the primary sequence repositories (GenBank, UniProt, and so on) can be kept current.
One of the fundamental questions in phylogenomic inference of protein function is determining the evolutionary distance within which annotations may be transferred. Given the paucity of sequences with experimentally determined function, if annotations can only be transferred between orthologs (and are also restricted to annotations having experimental support), the vast majority of unknown sequences will remain without predicted function. Is this necessary or overkill? Analysis of different types of 'function' associated with proteins show that some types of attributes (for example, catalytic activity) persist over large evolutionary distances, while in other cases (for example, substrate specificity), functions can diverge extremely rapidly. Moreover, the degree to which different types of function persist over evolutionary distance can vary from one family to another. One intriguing possibility for the next generation of phylogenomic inference methods involves identifying attribute-specific evolutionary distances over which attributes may percolate.
Finally, assessing annotation accuracy is a very labor-intensive practice. Biological curators can spend days analyzing and annotating a single gene; to do this in high-throughput for thousands of sequences is clearly not feasible. An additional complication is that definitions of molecular 'function' or 'subfunction' are not at all standardized within biology. Instead, some biologists use the term 'function' very specifically (see for example, ) while others may use the term more loosely. Assessing annotation accuracy, and comparing the relative effectiveness of different function prediction protocols, also requires judgment calls regarding definitions of correctness. An annotation may be technically correct, but at such a high level that it is minimally helpful. For instance, a novel gene may be labeled as 'putative membrane protein', 'putative GPCR' and 'putative chemokine receptor'. If experimental studies show the protein to be a chemokine receptor, then only the third annotation would be particularly helpful to biologists, although all annotations would be technically correct. In other cases, an annotation may be technically incorrect, although quite close, due to annotation transfer from a paralog with a slightly different functional specificity (for example, Serotonin receptor type 1 versus Serotonin receptor type 2). Critically, there is also no community-accepted benchmark dataset or scoring function to evaluate methods of protein functional classification. These need to be developed to enable computational biologists to determine what types of inference methods are robust under what conditions, and where our methods fail. Efforts to develop true de novo function prediction efforts, analogous to the biennial CASP protein structure prediction experiments, are underway , and are likely to play an important role in improving our understanding of method accuracy in this important area.
This work was supported by a Presidential Early Career Award for Scientists and Engineers (PECASE) from the National Science Foundation, and by an R01 from the National Human Genome Research Institute of the NIH. Neither funding source was involved in: the study design; the collection, analysis, and interpretation of data; the writing of the manuscript; or the decision to submit the manuscript for publication. The authors wish to thank several anonymous reviewers for very helpful comments and suggestions.
- Bork P, Koonin EV: Predicting functions from protein sequences - where are the bottlenecks?. Nat Genet. 1998, 18: 313-318. 10.1038/ng0498-313.PubMedView ArticleGoogle Scholar
- Galperin MY, Koonin EV: Sources of systematic error in functional annotation of genomes: domain rearrangement, non-orthologous gene displacement and operon disruption. In Silico Biol. 1998, 1: 55-67.PubMedGoogle Scholar
- Gerlt JA, Babbitt PC: Can sequence determine function?. Genome Biol. 2000, 1: REVIEWS0005-10.1186/gb-2000-1-5-reviews0005.PubMedPubMed CentralView ArticleGoogle Scholar
- Fitch WM: Distinguishing homologous from analogous proteins. Syst Zool. 1970, 19: 99-113. 10.2307/2412448.PubMedView ArticleGoogle Scholar
- Kaessmann H, Zollner S, Nekrutenko A, Li WH: Signatures of domain shuffling in the human genome. Genome Res. 2002, 12: 1642-1650. 10.1101/gr.520702.PubMedPubMed CentralView ArticleGoogle Scholar
- Rajalingam R, Parham P, Abi-Rached L: Domain shuffling has been the main mechanism forming new hominoid killer cell Ig-like receptors. J Immunol. 2004, 172: 356-369.PubMedView ArticleGoogle Scholar
- Brenner SE: Errors in genome annotation. Trends Genet. 1999, 15: 132-133. 10.1016/S0168-9525(99)01706-0.PubMedView ArticleGoogle Scholar
- Gilks WR, Audit B, De Angelis D, Tsoka S, Ouzounis CA: Modeling the percolation of annotation errors in a database of protein sequences. Bioinformatics. 2002, 18: 1641-1649. 10.1093/bioinformatics/18.12.1641.PubMedView ArticleGoogle Scholar
- Gilks WR, Audit B, de Angelis D, Tsoka S, Ouzounis CA: Percolation of annotation errors through hierarchically structured protein sequence databases. Math Biosci. 2005, 193: 223-234. 10.1016/j.mbs.2004.08.001.PubMedView ArticleGoogle Scholar
- Altschul SF, Madden TL, Schaffer AA, Zhang J, Zhang Z, Miller W, Lipman DJ: Gapped BLAST and PSI-BLAST: a new generation of protein database search programs. Nucleic Acids Res. 1997, 25: 3389-3402. 10.1093/nar/25.17.3389.PubMedPubMed CentralView ArticleGoogle Scholar
- PhyloFacts Resource. [http://phylogenomics.berkeley.edu/phylofacts/]
- Moult J, Fidelis K, Zemla A, Hubbard T: Critical assessment of methods of protein structure prediction (CASP)-round V. Proteins. 2003, 53 (Suppl 6): 334-339. 10.1002/prot.10556.PubMedView ArticleGoogle Scholar
- Bradley P, Chivian D, Meiler J, Misura KM, Rohl CA, Schief WR, Wedemeyer WJ, Schueler-Furman O, Murphy P, Schonbrun J, et al: Rosetta predictions in CASP5: successes, failures, and prospects for complete automation. Proteins. 2003, 53 (Suppl 6): 457-468. 10.1002/prot.10552.PubMedView ArticleGoogle Scholar
- Pellegrini M, Marcotte EM, Thompson MJ, Eisenberg D, Yeates TO: Assigning protein functions by comparative genome analysis: protein phylogenetic profiles. Proc Natl Acad Sci USA. 1999, 96: 4285-4288. 10.1073/pnas.96.8.4285.PubMedPubMed CentralView ArticleGoogle Scholar
- Marcotte EM, Pellegrini M, Ng HL, Rice DW, Yeates TO, Eisenberg D: Detecting protein function and protein-protein interactions from genome sequences. Science. 1999, 285: 751-753. 10.1126/science.285.5428.751.PubMedView ArticleGoogle Scholar
- Pearson WR, Lipman DJ: Improved tools for biological sequence comparison. Proc Natl Acad Sci USA. 1988, 85: 2444-2448. 10.1073/pnas.85.8.2444.PubMedPubMed CentralView ArticleGoogle Scholar
- Park J, Karplus K, Barrett C, Hughey R, Haussler D, Hubbard T, Chothia C: Sequence comparisons using multiple sequences detect three times as many remote homologues as pairwise methods. J Mol Biol. 1998, 284: 1201-1210. 10.1006/jmbi.1998.2221.PubMedView ArticleGoogle Scholar
- Bateman A, Coin L, Durbin R, Finn RD, Hollich V, Griffiths-Jones S, Khanna A, Marshall M, Moxon S, Sonnhammer EL, et al: The Pfam protein families database. Nucleic Acids Res. 2004, D138-141. 10.1093/nar/gkh121. 32 DatabaseGoogle Scholar
- Schultz J, Milpetz F, Bork P, Ponting CP: SMART, a simple modular architecture research tool: identification of signaling domains. Proc Natl Acad Sci USA. 1998, 95: 5857-5864. 10.1073/pnas.95.11.5857.PubMedPubMed CentralView ArticleGoogle Scholar
- Gough J, Karplus K, Hughey R, Chothia C: Assignment of homology to genome sequences using a library of hidden Markov models that represent all proteins of known structure. J Mol Biol. 2001, 313: 903-919. 10.1006/jmbi.2001.5080.PubMedView ArticleGoogle Scholar
- Eisen JA: Phylogenomics: improving functional predictions for uncharacterized genes by evolutionary analysis. Genome Res. 1998, 8: 163-167.PubMedView ArticleGoogle Scholar
- Sjölander K: Phylogenomic inference of protein molecular function: advances and challenges. Bioinformatics. 2004, 20: 170-179. 10.1093/bioinformatics/bth021.PubMedView ArticleGoogle Scholar
- Brown D, Sjölander K: Functional classification using phylogenomic inference. PLoS Comput Biol. 2006, 2: e77-10.1371/journal.pcbi.0020077.PubMedPubMed CentralView ArticleGoogle Scholar
- Storm CE, Sonnhammer EL: Automated ortholog inference from phylogenetic trees and calculation of orthology reliability. Bioinformatics. 2002, 18: 92-99. 10.1093/bioinformatics/18.1.92.PubMedView ArticleGoogle Scholar
- Zmasek CM, Eddy SR: RIO: Analyzing proteomes by automated phylogenomics using resampled inference of orthologs. BMC Bioinformatics. 2002, 3: 14-10.1186/1471-2105-3-14.PubMedPubMed CentralView ArticleGoogle Scholar
- Engelhardt BE, Jordan MI, Muratore KE, Brenner SE: Protein molecular function prediction by Bayesian phylogenomics. PLoS Comput Biol. 2005, 1: e45-10.1371/journal.pcbi.0010045.PubMedPubMed CentralView ArticleGoogle Scholar
- Krogh A, Larsson B, von Heijne G, Sonnhammer EL: Predicting transmembrane protein topology with a hidden Markov model: application to complete genomes. J Mol Biol. 2001, 305: 567-580. 10.1006/jmbi.2000.4315.PubMedView ArticleGoogle Scholar
- Emanuelsson O, Nielsen H, Brunak S, von Heijne G: Predicting subcellular localization of proteins based on their N-terminal amino acid sequence. J Mol Biol. 2000, 300: 1005-1016. 10.1006/jmbi.2000.3903.PubMedView ArticleGoogle Scholar
- Kall L, Krogh A, Sonnhammer EL: A combined transmembrane topology and signal peptide prediction method. J Mol Biol. 2004, 338: 1027-1036. 10.1016/j.jmb.2004.03.016.PubMedView ArticleGoogle Scholar
- Chen CP, Kernytsky A, Rost B: Transmembrane helix predictions revisited. Protein Sci. 2002, 11: 2774-2791. 10.1110/ps.0214502.PubMedPubMed CentralView ArticleGoogle Scholar
- Ashburner M, Ball CA, Blake JA, Botstein D, Butler H, Cherry JM, Davis AP, Dolinski K, Dwight SS, Eppig JT, et al: Gene ontology: tool for the unification of biology. The Gene Ontology Consortium. Nat Genet. 2000, 25: 25-29. 10.1038/75556.PubMedPubMed CentralView ArticleGoogle Scholar
- Browse PhyloFacts Universe. [http://phylogenomics.berkeley.edu/cgi-bin/listbooks/listbooks.py]
- PhyloFacts Search. [http://phylogenomics.berkeley.edu/phylofacts/search.php]
- Brenner SE, Koehl P, Levitt M: The ASTRAL compendium for protein structure and sequence analysis. Nucleic Acids Res. 2000, 28: 254-256. 10.1093/nar/28.1.254.PubMedPubMed CentralView ArticleGoogle Scholar
- Berman HM, Westbrook J, Feng Z, Gilliland G, Bhat TN, Weissig H, Shindyalov IN, Bourne PE: The Protein Data Bank. Nucleic Acids Res. 2000, 28: 235-242. 10.1093/nar/28.1.235.PubMedPubMed CentralView ArticleGoogle Scholar
- Wheeler DL, Barrett T, Benson DA, Bryant SH, Canese K, Chetvernin V, Church DM, DiCuccio M, Edgar R, Federhen S, et al: Database resources of the National Center for Biotechnology Information. Nucleic Acids Res. 2006, D173-180. 10.1093/nar/gkj158. 34 DatabaseGoogle Scholar
- Karplus K, Karchin R, Barrett C, Tu S, Cline M, Diekhans M, Grate L, Casper J, Hughey R: What is the value added by human intervention in protein structure prediction?. Proteins. 2001, 45 (Suppl 5): 86-91. 10.1002/prot.10021.View ArticleGoogle Scholar
- Horn F, Weare J, Beukers MW, Horsch S, Bairoch A, Chen W, Edvardsen O, Campagne F, Vriend G: GPCRDB: an information system for G protein-coupled receptors. Nucleic Acids Res. 1998, 26: 275-279. 10.1093/nar/26.1.275.PubMedPubMed CentralView ArticleGoogle Scholar
- Mi H, Lazareva-Ulitsky B, Loo R, Kejariwal A, Vandergriff J, Rabkin S, Guo N, Muruganujan A, Doremieux O, Campbell MJ, et al: The PANTHER database of protein families, subfamilies, functions and pathways. Nucleic Acids Res. 2005, D284-288. 33 DatabaseGoogle Scholar
- Drysdale RA, Crosby MA: FlyBase: genes and gene models. Nucleic Acids Res. 2005, D390-395. 33 DatabaseGoogle Scholar
- Jaiswal P, Ni J, Yap I, Ware D, Spooner W, Youens-Clark K, Ren L, Liang C, Zhao W, Ratnapu K, et al: Gramene: a bird's eye view of cereal genomes. Nucleic Acids Res. 2006, D717-723. 10.1093/nar/gkj154. 34 DatabaseGoogle Scholar
- Kent WJ, Sugnet CW, Furey TS, Roskin KM, Pringle TH, Zahler AM, Haussler D: The human genome browser at UCSC. Genome Res. 2002, 12: 996-1006. 10.1101/gr.229102. Article published online before print in May 2002.PubMedPubMed CentralView ArticleGoogle Scholar
- Rhee SY, Beavis W, Berardini TZ, Chen G, Dixon D, Doyle A, Garcia-Hernandez M, Huala E, Lander G, Montoya M, et al: The Arabidopsis Information Resource (TAIR): a model organism database providing a centralized, curated gateway to Arabidopsis biology, research materials and community. Nucleic Acids Res. 2003, 31: 224-228. 10.1093/nar/gkg076.PubMedView ArticleGoogle Scholar
- Zmasek CM, Eddy SR: A simple algorithm to infer gene duplication and speciation events on a gene tree. Bioinformatics. 2001, 17: 821-828. 10.1093/bioinformatics/17.9.821.PubMedView ArticleGoogle Scholar
- Marchler-Bauer A, Anderson JB, Cherukuri PF, DeWeese-Scott C, Geer LY, Gwadz M, He S, Hurwitz DI, Jackson JD, Ke Z, et al: CDD: a Conserved Domain Database for protein classification. Nucleic Acids Res. 2005, D192-196. 33 DatabaseGoogle Scholar
- Koonin EV: Orthologs, paralogs, and evolutionary genomics. Annu Rev Genet. 2005, 39: 309-338. 10.1146/annurev.genet.39.073003.114725.PubMedView ArticleGoogle Scholar
- Wu CH, Apweiler R, Bairoch A, Natale DA, Barker WC, Boeckmann B, Ferro S, Gasteiger E, Huang H, Lopez R, et al: The Universal Protein Resource (UniProt): an expanding universe of protein information. Nucleic Acids Res. 2006, D187-191. 10.1093/nar/gkj161. 34 DatabaseGoogle Scholar
- Eisen JA, Sweder KS, Hanawalt PC: Evolution of the SNF2 family of proteins: subfamilies with distinct sequences and functions. Nucleic Acids Res. 1995, 23: 2715-2723.PubMedPubMed CentralView ArticleGoogle Scholar
- Shakhnovich BE, Harvey JM, Comeau S, Lorenz D, DeLisi C, Shakhnovich E: ELISA: structure-function inferences based on statistically significant and evolutionarily inspired observations. BMC Bioinformatics. 2003, 4: 34-10.1186/1471-2105-4-34.PubMedPubMed CentralView ArticleGoogle Scholar
- Levy ED, Ouzounis CA, Gilks WR, Audit B: Probabilistic annotation of protein sequences based on functional classifications. BMC Bioinformatics. 2005, 6: 302-10.1186/1471-2105-6-302.PubMedPubMed CentralView ArticleGoogle Scholar
- Friedberg I: Automated protein function prediction - the genomic challenge. Brief Bioinform. 2006, 7: 225-242. 10.1093/bib/bbl004.PubMedView ArticleGoogle Scholar
- Fritz-Laylin LK, Krishnamurthy N, Tor M, Sjölander KV, Jones JD: Phylogenomic analysis of the receptor-like proteins of rice and Arabidopsis. Plant Physiol. 2005, 138: 611-623. 10.1104/pp.104.054452.PubMedPubMed CentralView ArticleGoogle Scholar
- Chatterjee S, Han H, Rollins S, Cleveland T: Molecular cloning, characterization, and expression of a novel human neutral sphingomyelinase. J Biol Chem. 1999, 274: 37407-37412. 10.1074/jbc.274.52.37407.PubMedView ArticleGoogle Scholar
- Berghard A, Dryer L: A novel family of ancient vertebrate odorant receptors. J Neurobiol. 1998, 37: 383-392. 10.1002/(SICI)1097-4695(19981115)37:3<383::AID-NEU4>3.0.CO;2-D.PubMedView ArticleGoogle Scholar
- Flowerpower Online Server. [http://phylogenomics.berkeley.edu/flowerpower/]
- Edgar RC, Sjölander K: COACH: profile-profile alignment of protein families using hidden Markov models. Bioinformatics. 2004, 20: 1309-1318. 10.1093/bioinformatics/bth091.PubMedView ArticleGoogle Scholar
- Edgar RC: MUSCLE: a multiple sequence alignment method with reduced time and space complexity. BMC Bioinformatics. 2004, 5: 113-10.1186/1471-2105-5-113.PubMedPubMed CentralView ArticleGoogle Scholar
- Edgar RC, Sjölander K: SATCHMO: sequence alignment and tree construction using hidden Markov models. Bioinformatics. 2003, 19: 1404-1411. 10.1093/bioinformatics/btg158.PubMedView ArticleGoogle Scholar
- Clamp M, Cuff J, Searle SM, Barton GJ: The Jalview Java alignment editor. Bioinformatics. 2004, 20: 426-427. 10.1093/bioinformatics/btg430.PubMedView ArticleGoogle Scholar
- Henikoff S, Henikoff JG: Amino acid substitution matrices from protein blocks. Proc Natl Acad Sci USA. 1992, 89: 10915-10919. 10.1073/pnas.89.22.10915.PubMedPubMed CentralView ArticleGoogle Scholar
- Sjölander K: Phylogenetic inference in protein superfamilies: analysis of SH2 domains. Proc Int Conf Intell Syst Mol Biol. 1998, 6: 165-174.PubMedGoogle Scholar
- Sjölander K, Karplus K, Brown M, Hughey R, Krogh A, Mian IS, Haussler D: Dirichlet mixtures: a method for improved detection of weak but significant protein sequence homology. Comput Appl Biosci. 1996, 12: 327-345.PubMedGoogle Scholar
- SCI-PHY Online Server. [http://phylogenomics.berkeley.edu/SCI-PHY/]
- Karplus K, Sjölander K, Barrett C, Cline M, Haussler D, Hughey R, Holm L, Sander C: Predicting protein structure using hidden Markov models. Proteins. 1997, 134-139. 10.1002/(SICI)1097-0134(1997)1+<134::AID-PROT18>3.0.CO;2-P. Suppl 1Google Scholar
- Brown D, Krishnamurthy N, Dale JM, Christopher W, Sjölander K: Subfamily hmms in functional genomics. Pac Symp Biocomput. 2005, 322-333.Google Scholar
- PHYLIP (Phylogenetic Inference Package) version 3.6. [http://evolution.genetics.washington.edu/phylip.html]
- Jones DT, Taylor WR, Thornton JM: The rapid generation of mutation data matrices from protein sequences. Comput Appl Biosci. 1992, 8: 275-282.PubMedGoogle Scholar
- Guindon S, Gascuel O: A simple, fast, and accurate algorithm to estimate large phylogenies by maximum likelihood. Syst Biol. 2003, 52: 696-704. 10.1080/10635150390235520.PubMedView ArticleGoogle Scholar
- Swofford D: PAUP*. Phylogenetic Analysis Using Parsimony (*and Other Methods). 2002, Sunderland, Massachusetts: Sinauer Associates, Version 4Google Scholar
- Zmasek CM, Eddy SR: ATV: display and manipulation of annotated phylogenetic trees. Bioinformatics. 2001, 17: 383-384. 10.1093/bioinformatics/17.4.383.PubMedView ArticleGoogle Scholar
- JMOL. [http://jmol.sourceforge.net]
- Pieper U, Eswar N, Davis FP, Braberg H, Madhusudhan MS, Rossi A, Marti-Renom M, Karchin R, Webb BM, Eramian D, et al: MODBASE: a database of annotated comparative protein structure models and associated resources. Nucleic Acids Res. 2006, D291-295. 10.1093/nar/gkj059. 34 DatabaseGoogle Scholar
- Soro S, Tramontano A: The prediction of protein function at CASP6. Proteins. 2005, 61 (Suppl 7): 201-213. 10.1002/prot.20738.PubMedView ArticleGoogle Scholar
- Celera Genomics Panther Classification. [http://www.pantherdb.org]
- TIGRFams. [http://www.tigr.org/TIGRFAMs/]
- PFAM HMM Library at the Sanger Institute. [http://www.sanger.ac.uk/Software/Pfam/]
- SMART. [http://smart.embl-heidelberg.de/]
- InterPro. [http://www.ebi.ac.uk/interpro/]
- Superfamily. [http://supfam.org/SUPERFAMILY/]
- Hall TM: Structure and function of argonaute proteins. Structure. 2005, 13: 1403-1408. 10.1016/j.str.2005.08.005.PubMedView ArticleGoogle Scholar
- Parker JS, Roe SM, Barford D: Crystal structure of a PIWI protein suggests mechanisms for siRNA recognition and slicer activity. EMBO J. 2004, 23: 4727-4737. 10.1038/sj.emboj.7600488.PubMedPubMed CentralView ArticleGoogle Scholar
- Rivas FV, Tolia NH, Song JJ, Aragon JP, Liu J, Hannon GJ, Joshua-Tor L: Purified Argonaute2 and an siRNA form recombinant human RISC. Nat Struct Mol Biol. 2005, 12: 340-349. 10.1038/nsmb918.PubMedView ArticleGoogle Scholar
This article is published under license to BioMed Central Ltd. This is an open access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.