A systematic comparison of protein structure classifications: SCOP, CATH and FSSP.

BACKGROUND: Several methods of structural classification have been developed to introduce some order to the large amount of data present in the Protein Data Bank. Such methods facilitate structural comparisons and provide a greater understanding of structure and function. The most widely used and comprehensive databases are SCOP, CATH and FSSP, which represent three unique methods of classifying protein structures: purely manual, a combination of manual and automated, and purely automated, respectively. In order to develop reliable template libraries and benchmarks for protein-fold recognition, a systematic comparison of these databases has been carried out to determine their overall agreement in classifying protein structures. RESULTS: Approximately two-thirds of the protein chains in each database are common to all three databases. Despite employing different methods, and basing their systems on different rules of protein structure and taxonomy, SCOP, CATH and FSSP agree on the majority of their classifications. Discrepancies and inconsistencies are accounted for by a small number of explanations. Other interesting features have been identified, and various differences between manual and automatic classification methods are presented. CONCLUSIONS: Using these databases requires an understanding of the rules upon which they are based; each method offers certain advantages depending on the biological requirements and knowledge of the user. The degree of discrepancy between the systems also has an impact on reliability of prediction methods that employ these schemes as benchmarks. To generate accurate fold templates for threading, we extract information from a consensus database, encompassing agreements between SCOP, CATH and FSSP.     

Automated assignment of SCOP and CATH protein structure classifications from FSSP scores

We present an automated procedure to assign CATH and SCOP classifications to proteins whose FSSP score is available. CATH classification is assigned down to the topology level, and SCOP classification is assigned to the fold level. Because the FSSP database is updated weekly, this method makes it possible to update also CATH and SCOP with the same frequency. Our predictions have a nearly perfect success rate when ambiguous cases are discarded. These ambiguous cases are intrinsic in any protein structure classification that relies on structural information alone. Hence, we introduce the "twilight zone for structure classification." We further suggest that to resolve these ambiguous cases, other criteria of classification, based also on information about sequence and function, must be used.     

Exploring protein domain structure

The protein databank contains coordinates of over 10,000 protein structures, which constitute more than 25,000 structural domains in total. The investigation of protein structural, functional and evolutionary relationships is fundamental to many important fields in bioinformatics research, and will be crucial in determining the function of the human and other genomes.This review describes the SCOP and CATH databases of protein structure classification, which define, classify and annotate each domain in the protein databank. The hierarchical structure, use and annotation of the databases are explained. Other tools for exploring protein structure relationships are also described.     

DDOMAIN: Dividing structures into domains using a normalized domain-domain interaction profile

Dividing protein structures into domains is proven useful for more accurate structural and functional characterization of proteins. Here, we develop a method, called DDOMAIN, that divides structure into DOMAINs using a normalized contact-based domain-domain interaction profile. Results of DDOMAIN are compared to AUTHORS annotations (domain definitions are given by the authors who solved protein structures), as well as to popular SCOP and CATH annotations by human experts and automatic programs. DDOMAIN's automatic annotations are most consistent with the AUTHORS annotations (90% agreement in number of domains and 88% agreement in both number of domains and at least 85% overlap in domain assignment of residues) if its three adjustable parameters are trained by the AUTHORS annotations. By comparison, the agreement is 83% (81% with at least 85% overlap criterion) between SCOP-trained DDOMAIN and SCOP annotations and 77% (73%) between CATH-trained DDOMAIN and CATH annotations. The agreement between DDOMAIN and AUTHORS annotations goes beyond single-domain proteins (97%, 82%, and 56% for single-, two-, and three-domain proteins, respectively). For an "easy" data set of proteins whose CATH and SCOP annotations agree with each other in number of domains, the agreement is 90% (89%) between "easy-set"-trained DDOMAIN and CATH/SCOP annotations. The consistency between SCOP-trained DDOMAIN and SCOP annotations is superior to two other recently developed, SCOP-trained, automatic methods PDP (protein domain parser), and DomainParser 2. We also tested a simple consensus method made of PDP, DomainParser 2, and DDOMAIN and a different version of DDOMAIN based on a more sophisticated statistical energy function. The DDOMAIN server and its executable are available in the services section on http://sparks.informatics.iupui.edu.     

A comparison of SCOP and CATH with respect to domain-domain interactions

The analysis and prediction of protein-protein interaction sites from structural data are restricted by the limited availability of structural complexes that represent the complete protein-protein interaction space. The domain classification schemes CATH and SCOP are normally used independently in the analysis and prediction of protein domain-domain interactions. In this article, the effect of different domain classification schemes on the number and type of domain-domain interactions observed in structural data is systematically evaluated for the SCOP and CATH hierarchies. Although there is a large overlap in domain assignments between SCOP and CATH, 23.6% of CATH interfaces had no SCOP equivalent and 37.3% of SCOP interfaces had no CATH equivalent in a nonredundant set. Therefore, combining both classifications gives an increase of between 23.6 and 37.3% in domain-domain interfaces. It is suggested that if possible, both domain classification schemes should be used together, but if only one is selected, SCOP provides better coverage than CATH. Employing both SCOP and CATH reduces the false negative rate of predictive methods, which employ homology matching to structural data to predict protein-protein interaction by an estimated 6.5%.     

Assessing annotation transfer for genomics: quantifying the relations between protein sequence, structure and function through traditional and probabilistic scores

Measuring in a quantitative, statistical sense the degree to which structural and functional information can be "transferred" between pairs of related protein sequences at various levels of similarity is an essential prerequisite for robust genome annotation. To this end, we performed pairwise sequence, structure and function comparisons on approximately 30,000 pairs of protein domains with known structure and function. Our domain pairs, which are constructed according to the SCOP fold classification, range in similarity from just sharing a fold, to being nearly identical. Our results show that traditional scores for sequence and structure similarity have the same basic exponential relationship as observed previously, with structural divergence, measured in RMS, being exponentially related to sequence divergence, measured in percent identity. However, as the scale of our survey is much larger than any previous investigations, our results have greater statistical weight and precision. We have been able to express the relationship of sequence and structure similarity using more "modern scores," such as Smith-Waterman alignment scores and probabilistic P-values for both sequence and structure comparison. These modern scores address some of the problems with traditional scores, such as determining a conserved core and correcting for length dependency; they enable us to phrase the sequence-structure relationship in more precise and accurate terms. We found that the basic exponential sequence-structure relationship is very general: the same essential relationship is found in the different secondary-structure classes and is evident in all the scoring schemes. To relate function to sequence and structure we assigned various levels of functional similarity to the domain pairs, based on a simple functional classification scheme. This scheme was constructed by combining and augmenting annotations in the enzyme and fly functional classifications and comparing subsets of these to the Escherichia coli and yeast classifications. We found sigmoidal relationships between similarity in function and sequence, with clear thresholds for different levels of functional conservation. For pairs of domains that share the same fold, precise function appears to be conserved down to approximately 40 % sequence identity, whereas broad functional class is conserved to approximately 25 %. Interestingly, percent identity is more effective at quantifying functional conservation than the more modern scores (e.g. P-values). Results of all the pairwise comparisons and our combined functional classification scheme for protein structures can be accessed from a web database at http://bioinfo.mbb.yale.edu/alignCopyright 2000 Academic Press.     

Current status of membrane protein structure classification

For over 2 decades, continuous efforts to organize the jungle of available protein structures have been underway. Although a number of discrepancies between different classification approaches for soluble proteins have been reported, the classification of membrane proteins has so far not been comparatively studied because of the limited amount of available structural data. Here, we present an analysis of α‐helical membrane protein classification in the SCOP and CATH databases. In the current set of 63 α‐helical membrane protein chains having between 1 and 13 transmembrane helices, we observed a number of differently classified proteins both regarding their domain and fold assignment. The majority of all discrepancies affect single transmembrane helix, two helix hairpin, and four helix bundle domains, while domains with more than five helices are mostly classified consistently between SCOP and CATH. It thus appears that the structural constraints imposed by the lipid bilayer complicate the classification of membrane proteins with only few membrane‐spanning regions. This problem seems to be specific for membrane proteins as soluble four helix bundles, not restrained by the membrane, are more consistently classified by SCOP and CATH. Our findings indicate that the structural space of small membrane helix bundles is highly continuous such that even minor differences in individual classification procedures may lead to a significantly different classification. Membrane proteins with few helices and limited structural diversity only seem to be reasonably classifiable if the definition of a fold is adapted to include more fine‐grained structural features such as helix–helix interactions and reentrant regions. Proteins 2010. © 2010 Wiley‐Liss, Inc.     

Toward consistent assignment of structural domains in proteins

The assignment of protein domains from three-dimensional structure is critically important in understanding protein evolution and function, yet little quality assurance has been performed. Here, the differences in the assignment of structural domains are evaluated using six common assignment methods. Three human expert methods (AUTHORS (authors' annotation), CATH and SCOP) and three fully automated methods (DALI, DomainParser and PDP) are investigated by analysis of individual methods against the author's assignment as well as analysis based on the consensus among groups of methods (only expert, only automatic, combined). The results demonstrate that caution is recommended in using current domain assignments, and indicates where additional work is needed. Specifically, the major factors responsible for conflicting domain assignments between methods, both experts and automatic, are: (1) the definition of very small domains; (2) splitting secondary structures between domains; (3) the size and number of discontinuous domains; (4) closely packed or convoluted domain-domain interfaces; (5) structures with large and complex architectures; and (6) the level of significance placed upon structural, functional and evolutionary concepts in considering structural domain definitions. A web-based resource that focuses on the results of benchmarking and the analysis of domain assignments is available at     

Recognizing the fold of a protein structure

This paper reports a graph-theoretic program, GRATH, that rapidly, and accurately, matches a novel structure against a library of domain structures to find the most similar ones. GRATH generates distributions of scores by comparing the novel domain against the different types of folds that have been classified previously in the CATH database of structural domains. GRATH uses a measure of similarity that details the geometric information, number of secondary structures and number of residues within secondary structures, that any two protein structures share. Although GRATH builds on well established approaches for secondary structure comparison, a novel scoring scheme has been introduced to allow ranking of any matches identified by the algorithm. More importantly, we have benchmarked the algorithm using a large dataset of 1702 non-redundant structures from the CATH database which have already been classified into fold groups, with manual validation. This has facilitated introduction of further constraints, optimization of parameters and identification of reliable thresholds for fold identification. Following these benchmarking trials, the correct fold can be identified with the top score with a frequency of 90%. It is identified within the ten most likely assignments with a frequency of 98%. GRATH has been implemented to use via a server (http://www.biochem.ucl.ac.uk/cgi-bin/cath/Grath.pl). GRATH's speed and accuracy means that it can be used as a reliable front-end filter for the more accurate, but computationally expensive, residue based structure comparison algorithm SSAP, currently used to classify domain structures in the CATH database. With an increasing number of structures being solved by the structural genomics initiatives, the GRATH server also provides an essential resource for determining whether newly determined structures are related to any known structures from which functional properties may be inferred.     

FuzzyART neural network for protein classification

One of the major research directions in bioinformatics is that of predicting the protein superfamily in large databases and classifying a given set of protein domains into superfamilies. The classification reflects the structural, evolutionary and functional relatedness. These relationships are embodied in hierarchical classification such as Structural Classification of Protein (SCOP), which is manually curated. Such classification is essential for the structural and functional analysis of proteins. Yet, a large number of proteins remain unclassified. We have proposed an unsupervised machine-learning FuzzyART neural network algorithm to classify a given set of proteins into SCOP superfamilies. The proposed method is fast learning and uses an atypical non-linear pattern recognition technique. In this approach, we have constructed a similarity matrix from p-values of BLAST all-against-all, trained the network with FuzzyART unsupervised learning algorithm using the similarity matrix as input vectors and finally the trained network offers SCOP superfamily level classification. In this experiment, we have evaluated the performance of our method with existing techniques on six different datasets. We have shown that the trained network is able to classify a given similarity matrix of a set of sequences into SCOP superfamilies at high classification accuracy.     

Progress of structural genomics initiatives: an analysis of solved target structures

The explosion in gene sequence data and technological breakthroughs in protein structure determination inspired the launch of structural genomics (SG) initiatives. An often stated goal of structural genomics is the high-throughput structural characterisation of all protein sequence families, with the long-term hope of significantly impacting on the life sciences, biotechnology and drug discovery. Here, we present a comprehensive analysis of solved SG targets to assess progress of these initiatives. Eleven consortia have contributed 316 non-redundant entries and 323 protein chains to the Protein Data Bank (PDB), and 459 and 393 domains to the CATH and SCOP structure classifications, respectively. The quality and size of these proteins are comparable to those solved in traditional structural biology and, despite huge scope for duplicated efforts, only 14% of targets have a close homologue (>/=30% sequence identity) solved by another consortium. Analysis of CATH and SCOP revealed the significant contribution that structural genomics is making to the coverage of superfamilies and folds. A total of 67% of SG domains in CATH are unique, lacking an already characterised close homologue in the PDB, whereas only 21% of non-SG domains are unique. For 29% of domains, structure determination revealed a remote evolutionary relationship not apparent from sequence, and 19% and 11% contributed new superfamilies and folds. The secondary structure class, fold and superfamily distributions of this dataset reflect those of the genomes. The domains fall into 172 different folds and 259 superfamilies in CATH but the distribution is highly skewed. The most populous of these are those that recur most frequently in the genomes. Whilst 11% of superfamilies are bacteria-specific, most are common to all three superkingdoms of life and together the 316 PDB entries have provided new and reliable homology models for 9287 non-redundant gene sequences in 206 completely sequenced genomes. From the perspective of this analysis, it appears that structural genomics is on track to be a success, and it is hoped that this work will inform future directions of the field.     

Structural classification of thioredoxin-like fold proteins

Protein structure classification is necessary to comprehend the rapidly growing structural data for better understanding of protein evolution and sequence-structure-function relationships. Thioredoxins are important proteins that ubiquitously regulate cellular redox status and various other crucial functions. We define the thioredoxin-like fold using the structure consensus of thioredoxin homologs and consider all circular permutations of the fold. The search for thioredoxin-like fold proteins in the PDB database identified 723 protein domains. These domains are grouped into eleven evolutionary families based on combined sequence, structural, and functional evidence. Analysis of the protein-ligand structure complexes reveals two major active site locations for the thioredoxin-like proteins. Comparison to existing structure classifications reveals that our thioredoxin-like fold group is broader and more inclusive, unifying proteins from five SCOP folds, five CATH topologies and seven DALI domain dictionary globular folding topologies. Considering these structurally similar domains together sheds new light on the relationships between sequence, structure, function and evolution of thioredoxins.     

Cognate ligand domain mapping for enzymes

Here, we present an automatic assignment of potential cognate ligands to domains of enzymes in the CATH and SCOP protein domain classifications on the basis of structural data available in the wwPDB. This procedure involves two steps; firstly, we assign the binding of particular ligands to particular domains; secondly, we compare the chemical similarity of the PDB ligands to ligands in KEGG in order to assign cognate ligands. We find that use of the Enzyme Commission (EC) numbers is necessary to enable efficient and accurate cognate ligand assignment. The PROCOGNATE database currently has cognate ligand mapping for 3277 (4118) protein structures and 351 (302) superfamilies, as described by the CATH and (SCOP) databases, respectively. We find that just under half of all ligands are only and always bound by a single domain, with 16% bound by more than one domain and the remainder of the ligands showing a variety of binding modes. This finding has implications for domain recombination and the evolution of new protein functions. Domain architecture or context is also found to affect substrate specificity of particular domains, and we discuss example cases. The most popular PDB ligands are all found to be generic components of crystallisation buffers, highlighting the non-cognate ligand problem inherent in the PDB. In contrast, the most popular cognate ligands are all found to be universal cellular currencies of reducing power and energy such as NADH, FADH2 and ATP, respectively, reflecting the fact that the vast majority of enzymatic reactions utilise one of these popular co-factors. These ligands all share a common adenine ribonucleotide moiety, suggesting that many different domain superfamilies have converged to bind this chemical framework.     

Functional fingerprints of folds: evidence for correlated structure-function evolution

Using structural similarity clustering of protein domains: protein domain universe graph (PDUG), and a hierarchical functional annotation: gene ontology (GO) as two evolutionary lenses, we find that each structural cluster (domain fold) exhibits a distribution of functions that is unique to it. These functional distributions are functional fingerprints that are specific to characteristic structural clusters and vary from cluster to cluster. Furthermore, as structural similarity threshold for domain clustering in the PDUG is relaxed we observe an influx of earlier-diverged domains into clusters. These domains join clusters without destroying the functional fingerprint. These results can be understood in light of a divergent evolution scenario that posits correlated divergence of structural and functional traits in protein domains from one or few progenitors.     

Comparison of sequence and structure-based datasets for nonredundant structural data mining

Structural data mining studies attempt to deduce general principles of protein structure from solved structures deposited in the protein data bank (PDB). The entire database is unsuitable for such studies because it is not representative of the ensemble of protein folds. Given that novel folds continue to be unearthed, some folds are currently unrepresented in the PDB while other folds are overrepresented. Overrepresentation can easily be avoided by filtering the dataset. PDB_SELECT is a well-used representative subset of the PDB that has been deduced by sequence comparison. Specifically, structures with sequences that exhibit a pairwise sequence identity above a threshold value are weeded from the dataset. Although length criteria for pairwise alignments have a structural basis, this automated method of pruning is essentially sequence-based and runs into problems in the twilight zone, possibly resulting in some folds being overrepresented. The value-added structure databases SCOP and CATH are also a potential source of a nonredundant dataset. Here we compare the sequence-derived dataset PDB_SELECT with the structural databases SCOP (Structural Classification Of Proteins) and CATH (Class-Architecture-Topology-Homology). We show that some folds remain overrepresented in the PDB_SELECT dataset while other folds are not represented at all. However, SCOP and CATH also have their own problems such as the labor-intensiveness of the update process and the problem of determining whether all folds are equally or sufficiently distant. We discuss areas where further work is required.     

Evolution of the beta-propeller fold

beta-Propellers are toroidal folds, in which repeated, four-stranded beta-meanders are arranged in a circular and slightly tilted fashion, like the blades of a propeller. They are found in all domains of life, with a strong preponderance among eukaryotes. Propellers show considerable sequence diversity and are classified into six separate structural groups by the SCOP and CATH databases. Despite this diversity, they often show similarities across groups, not only in structure but also in sequence, raising the possibility of a common origin. In agreement with this hypothesis, most propellers group together in a cluster map of all-beta folds generated by sequence similarity, because of numerous pairwise matches, many of which are individually nonsignificant. In total, 45 of 60 propellers in the SCOP25 database, covering four SCOP folds, are clustered in this group and analysis with sensitive sequence comparison methods shows that they are similar at a level indicative of homology. Two mechanisms appear to contribute to the evolution of beta-propellers: amplification from single blades and subsequent functional differentiation. The observation of propellers with nearly identical blades in genomic sequences show that these mechanisms are still operating today.     

Towards an automatic classification of protein structural domains based on structural similarity

Background  

Formal classification of a large collection of protein structures aids the understanding of evolutionary relationships among them. Classifications involving manual steps, such as SCOP and CATH, face the challenge of increasing volume of available structures. Automatic methods such as FSSP or Dali Domain Dictionary, yield divergent classifications, for reasons not yet fully investigated. One possible reason is that the pairwise similarity scores used in automatic classification do not adequately reflect the judgments made in manual classification. Another possibility is the difference between manual and automatic classification procedures. We explore the degree to which these two factors might affect the final classification.     

Identification of homology in protein structure classification.

Structural biology and structural genomics are expected to produce many three-dimensional protein structures in the near future. Each new structure raises questions about its function and evolution. Correct functional and evolutionary classification of a new structure is difficult for distantly related proteins and error-prone using simple statistical scores based on sequence or structure similarity. Here we present an accurate numerical method for the identification of evolutionary relationships (homology). The method is based on the principle that natural selection maintains structural and functional continuity within a diverging protein family. The problem of different rates of structural divergence between different families is solved by first using structural similarities to produce a global map of folds in protein space and then further subdividing fold neighborhoods into superfamilies based on functional similarities. In a validation test against a classification by human experts (SCOP), 77% of homologous pairs were identified with 92% reliability. The method is fully automated, allowing fast, self-consistent and complete classification of large numbers of protein structures. In particular, the discrimination between analogy and homology of close structural neighbors will lead to functional predictions while avoiding overprediction.     

The relationship between protein structure and function: a comprehensive survey with application to the yeast genome.

For most proteins in the genome databases, function is predicted via sequence comparison. In spite of the popularity of this approach, the extent to which it can be reliably applied is unknown. We address this issue by systematically investigating the relationship between protein function and structure. We focus initially on enzymes functionally classified by the Enzyme Commission (EC) and relate these to by structurally classified domains the SCOP database. We find that the major SCOP fold classes have different propensities to carry out certain broad categories of functions. For instance, alpha/beta folds are disproportionately associated with enzymes, especially transferases and hydrolases, and all-alpha and small folds with non-enzymes, while alpha+beta folds have an equal tendency either way. These observations for the database overall are largely true for specific genomes. We focus, in particular, on yeast, analyzing it with many classifications in addition to SCOP and EC (i.e. COGs, CATH, MIPS), and find clear tendencies for fold-function association, across a broad spectrum of functions. Analysis with the COGs scheme also suggests that the functions of the most ancient proteins are more evenly distributed among different structural classes than those of more modern ones. For the database overall, we identify the most versatile functions, i.e. those that are associated with the most folds, and the most versatile folds, associated with the most functions. The two most versatile enzymatic functions (hydro-lyases and O-glycosyl glucosidases) are associated with seven folds each. The five most versatile folds (TIM-barrel, Rossmann, ferredoxin, alpha-beta hydrolase, and P-loop NTP hydrolase) are all mixed alpha-beta structures. They stand out as generic scaffolds, accommodating from six to as many as 16 functions (for the exceptional TIM-barrel). At the conclusion of our analysis we are able to construct a graph giving the chance that a functional annotation can be reliably transferred at different degrees of sequence and structural similarity. Supplemental information is available from http://bioinfo.mbb.yale.edu/genome/foldfunc++ +.