Prediction of protein domain boundaries from statistics of appearance of amino acid residues

A database of 452 two-domain proteins with less than 25% homology was constructed. One half of the database was used to obtain statistics on the appearance of amino acid residues at domain boundaries. Small and hydrophilic residues (proline, glycine, asparagine, glutamic acid, arginine, etc.) occurred more often at domain boundaries than in total proteins. Hydrophobic residues (tryptophan, methionine, phenylalanine, etc.) were rarer at domain boundaries than in total proteins. Probability scales of amino acid appearance in boundary-flanking regions were constructed with these statistics and used to predict the domain boundaries in proteins of the other half of the database. The probability scale obtained by averaging the appearance of amino acids over an 8-residue region (±4 residues from the real domain boundaries) yielded the best results: domain boundaries were predicted within 40 residues of the real boundary in 57% of proteins and within 20 residues of the real boundary in 41% of proteins. The probability scale was used to predict the domain boundaries in proteins with unknown structures (CASP6).     

Prediction of protein domain boundaries from sequence alone

We present here a simple approach to identify domain boundaries in proteins of an unknown three-dimensional structure. Our method is based on the hypothesis that a high-side chain entropy of a region in a protein chain must be compensated by a high-residue interaction energy within the region, which could correlate with a well-structured part of the globule, that is, with a domain unit. For protein domains, this means that the domain boundary is conditioned by amino acid residues with a small value of side chain entropy, which correlates with the side chain size. On the one hand, relatively high Ala and Gly content on the domain boundary results in high conformational entropy of the backbone chain between the domains. On the other hand, the presence of Pro residues leads to the formation of hinges for a relative orientation of domains. The method was applied to 646 proteins with two contiguous domains extracted from the SCOP database with a success rate of 63%. We also report the prediction of domain boundaries for CASP5 targets obtained with the same method.     

Domain size distributions can predict domain boundaries

MOTIVATION: The sizes of protein domains observed in the 3D-structure database follow a surprisingly narrow distribution. Structural domains are furthermore formed from a single-chain continuous segment in over 80% of instances. These observations imply that some choices of domain boundaries on an otherwise uncharacterized sequence are more likely than others, based solely on the size and segment number of predicted domains. This property might be used to guess the locations of protein domain boundaries. RESULTS: To test this possibility we enumerate putative domain boundaries and calculate their relative likelihood under a probability model that considers only the size and segment number of predicted domains. We ask, in a cross-validated test using sequences with known 3D structure, whether the most likely guesses agree with the observed domain structure. We find that domain boundary predictions are surprisingly successful for sequences up to 400 residues long and that guessing domain boundaries in this way can improve the sensitivity of threading analysis.     

Armadillo: domain boundary prediction by amino acid composition

The identification and annotation of protein domains provides a critical step in the accurate determination of molecular function. Both computational and experimental methods of protein structure determination may be deterred by large multi-domain proteins or flexible linker regions. Knowledge of domains and their boundaries may reduce the experimental cost of protein structure determination by allowing researchers to work on a set of smaller and possibly more successful alternatives. Current domain prediction methods often rely on sequence similarity to conserved domains and as such are poorly suited to detect domain structure in poorly conserved or orphan proteins. We present here a simple computational method to identify protein domain linkers and their boundaries from sequence information alone. Our domain predictor, Armadillo (http://armadillo.blueprint.org), uses any amino acid index to convert a protein sequence to a smoothed numeric profile from which domains and domain boundaries may be predicted. We derived an amino acid index called the domain linker propensity index (DLI) from the amino acid composition of domain linkers using a non-redundant structure dataset. The index indicates that Pro and Gly show a propensity for linker residues while small hydrophobic residues do not. Armadillo predicts domain linker boundaries from Z-score distributions and obtains 35% sensitivity with DLI in a two-domain, single-linker dataset (within +/-20 residues from linker). The combination of DLI and an entropy-based amino acid index increases the overall Armadillo sensitivity to 56% for two domain proteins. Moreover, Armadillo achieves 37% sensitivity for multi-domain proteins, surpassing most other prediction methods. Armadillo provides a simple, but effective method by which prediction of domain boundaries can be obtained with reasonable sensitivity. Armadillo should prove to be a valuable tool for rapidly delineating protein domains in poorly conserved proteins or those with no sequence neighbors. As a first-line predictor, domain meta-predictors could yield improved results with Armadillo predictions.     

FIEFDom: a transparent domain boundary recognition system using a fuzzy mean operator

Protein domain prediction is often the preliminary step in both experimental and computational protein research. Here we present a new method to predict the domain boundaries of a multidomain protein from its amino acid sequence using a fuzzy mean operator. Using the nr-sequence database together with a reference protein set (RPS) containing known domain boundaries, the operator is used to assign a likelihood value for each residue of the query sequence as belonging to a domain boundary. This procedure robustly identifies contiguous boundary regions. For a dataset with a maximum sequence identity of 30%, the average domain prediction accuracy of our method is 97% for one domain proteins and 58% for multidomain proteins. The presented model is capable of using new sequence/structure information without re-parameterization after each RPS update. When tested on a current database using a four year old RPS and on a database that contains different domain definitions than those used to train the models, our method consistently yielded the same accuracy while two other published methods did not. A comparison with other domain prediction methods used in the CASP7 competition indicates that our method performs better than existing sequence-based methods.     

A numerical measure of amino acid residues similarity based on the analysis of their surroundings in natural protein sequences

A measure of similarity between amino acid residues based on the analysis of the surroundings of each residue in primary structures of native proteins is proposed. The statistical data used for this purpose were obtained from the analysis of 168,808 protein sequences, which comprise the Protein Identification Research database (release 63). Using various threshold values of the proposed measure, amino acid residues were classified into several groups. The classification elaborated differs essentially from groupings previously used. The numerical measure of amino acid residues similarity can be used in site-directed mutagenesis studies for the prediction of probability of local spatial rearrangements in proteins.     

Sequence-based prediction of protein domains

Guessing the boundaries of structural domains has been an important and challenging problem in experimental and computational structural biology. Predictions were based on intuition, biochemical properties, statistics, sequence homology and other aspects of predicted protein structure. Here, we introduced CHOPnet, a de novo method that predicts structural domains in the absence of homology to known domains. Our method was based on neural networks and relied exclusively on information available for all proteins. Evaluating sustained performance through rigorous cross-validation on proteins of known structure, we correctly predicted the number of domains in 69% of all proteins. For 50% of the two-domain proteins the centre of the predicted boundary was closer than 20 residues to the boundary assigned from three-dimensional (3D) structures; this was about eight percentage points better than predictions by ‘equal split’. Our results appeared to compare favourably with those from previously published methods. CHOPnet may be useful to restrict the experimental testing of different fragments for structure determination in the context of structural genomics.     

Identification of domains and domain interface residues in multidomain proteins from graph spectral method

We present a novel method for the identification of structural domains and domain interface residues in proteins by graph spectral method. This method converts the three-dimensional structure of the protein into a graph by using atomic coordinates from the PDB file. Domain definitions are obtained by constructing either a protein backbone graph or a protein side-chain graph. The graph is constructed based on the interactions between amino acid residues in the three-dimensional structure of the proteins. The spectral parameters of such a graph contain information regarding the domains and subdomains in the protein structure. This is based on the fact that the interactions among amino acids are higher within a domain than across domains. This is evident in the spectra of the protein backbone and the side-chain graphs, thus differentiating the structural domains from one another. Further, residues that occur at the interface of two domains can also be easily identified from the spectra. This method is simple, elegant, and robust. Moreover, a single numeric computation yields both the domain definitions and the interface residues.     

Statistical analysis of unstructured amino acid residues in protein structures

We have performed a statistical analysis of unstructured amino acid residues in protein structures available in the databank of protein structures. Data on the occurrence of disordered regions at the ends and in the middle part of protein chains have been obtained: in the regions near the ends (at distance less than 30 residues from the N- or C-terminus), there are 66% of unstructured residues (38% are near the N-terminus and 28% are near the C-terminus), although these terminal regions include only 23% of the amino acid residues. The frequencies of occurrence of unstructured residues have been calculated for each of 20 types in different positions in the protein chain. It has been shown that relative frequencies of occurrence of unstructured residues of 20 types at the termini of protein chains differ from the ones in the middle part of the protein chain; amino acid residues of the same type have different probabilities to be unstructured in the terminal regions and in the middle part of the protein chain. The obtained frequencies of occurrence of unstructured residues in the middle part of the protein chain have been used as a scale for predicting disordered regions from amino acid sequence using the method (FoldUnfold) previously developed by us. This scale of frequencies of occurrence of unstructured residues correlates with the contact scale (previously developed by us and used for the same purpose) at a level of 95%. Testing the new scale on a database of 427 unstructured proteins and 559 completely structured proteins has shown that this scale can be successfully used for the prediction of disordered regions in protein chains.     

Using First Passage Statistics to Extract Environmentally Dependent Amino Acid Correlations

In this work, we study the first passage statistics of amino acid primary sequences, that is the probability of observing an amino acid for the first time at a certain number of residues away from a fixed amino acid. By using this rich mathematical framework, we are able to capture the background distribution for an organism, and infer lengths at which the first passage has a probability that differs from what is expected. While many features of an organism''s genome are due to natural selection, others are related to amino acid chemistry and the environment in which an organism lives, constraining the randomness of genomes upon which selection can further act. We therefore use this approach to infer amino acid correlations, and then study how these correlations vary across a wide range of organisms under a wide range of optimal growth temperatures. We find a nearly universal exponential background distribution, consistent with the idea that most amino acids are globally uncorrelated from other amino acids in genomes. When we are able to extract significant correlations, these correlations are reliably dependent on optimal growth temperature, across phylogenetic boundaries. Some of the correlations we extract, such as the enhanced probability of finding, for the first time, a cysteine three residues away from a cysteine or glutamic acid two residues away from an arginine, likely relate to thermal stability. However, other correlations, likely appearing on alpha helical surfaces, have a less clear physiochemical interpretation and may relate to thermal stability or unusual metabolic properties of organisms that live in a high temperature environment.     

Computer-based search for steroid and DNA binding sites on estrogen and glucocorticoid receptors

The primary amino acid sequences of proteins that are receptors for estrogen, glucocorticoids, and ouabain were compared with each other using computer programs designed to detect and quantify similarities between proteins. Three regions of similarity between the estrogen receptor (ER) and the glucocorticoid receptor (GR) were identified. On the ER, residues 173-250, 323-395, and 426-458 are similar to residues 409-486, 540-612, and 644-676, respectively, on the GR. The ALIGN computer analysis of these segments on the ER and the GR gave comparison scores that were 16.8, 13.7, and 6.8 standard deviations higher, respectively, than that obtained with a comparison of randomized sequences of these proteins. The probability of getting these scores by chance is less than 10(-60), 10(-40), and 10(-11), respectively. Others have proposed that the segment on the ER and GR that is nearest their amino terminus (e.g. residues 173-250 of the ER) is part of their DNA binding domain and that the other two similar segments on each receptor, which are closer to their carboxy terminus, are part of their steroid binding domain. Here, we present evidence to support both of these hypotheses. First, an Align computer analysis indicates that residues 323-395 of the ER and residues 570-612 of the GR contain a region that is similar to a part of the alpha-subunit of the (Na+ + K+)ATPase that is hypothesized to bind the steroid ouabain. This similarity provides additional support for the proposed location of the steroid binding site on the ER, GR, and (Na+ + K+)ATPase. Second, a computer search of the protein sequence database revealed that protamine, a DNA binding protein, has some similarity to residues 255-281 of the ER, which are thought to be part of the DNA binding domain in the ER. Further, we find that residues 276-281 of the ER contain a structure that has been found at the nucleotide binding domain of some protein kinases. If this region on the ER binds ATP, then it may be involved in phosphorylation/dephosphorylation of the ER, which is thought to be important in its mechanism of action.     

Identification and analysis of novel amino acid sequence repeats and domains in Pyrobaculum aerophilum using computational tools

We have identified four repeats and five domains that are novel in proteins encoded by the Pyrobaculum aerophilum str. IM2 proteome using automated in silico methods. A "repeat" corresponds to a region comprising less than 55 amino acid residues that occurs more than once in the protein sequence and sometimes present in tandem. A "domain" corresponds to a conserved region comprising greater than 55 amino acid residues and may be present as single or multiple copies in the protein sequence. These correspond to (1) 85 amino acid residues AAG domain, (2) 72 amino acid residues GFGN domain, (3) 43 amino acid residues KGG repeat, (4) 25 amino acid residues RWE repeat, (5) 25 amino acid residues RID repeat, (6) 108 amino acid residues NDFA domain, (7) 140 amino acid residues VxY domain, (8) 35 amino acid residues LLPN repeat and (9) 98 amino acid residues GxY domain. A repeat or domain is characterized by specific conserved sequence motifs. We discuss the presence of these repeats and domains in proteins from other genomes and their probable secondary structure.     

Prediction of the location of structural domains in globular proteins

The location of structural domains in proteins is predicted from the amino acid sequence, based on the analysis of a computed contact map for the protein, the average distance map (ADM). Interactions between residues i and j in a protein are subdivided into several ranges, according to the separation |i-j| in the amino acid sequence. Within each range, average spatial distances between every pair of amino acid residues are computed from a data base of known protein structures. Infrequently occurring pairs are omitted as being statistically insignificant. The average distances are used to construct a predicted ADM. The ADM is analyzed for the occurrence of regions with high densities of contacts (compact regions). Locations of rapid changes of density between various parts of the map are determined by the use of scanning plots of contact densities. These locations serve to pinpoint the distribution of compact regions. This distribution, in turn, is used to predict boundaries of domains in the protein. The technique provides an objective method for the location of domains both on a contact map derived from a known three-dimensional protein structure, the real distance map (RDM), and on an ADM. While most other published methods for the identification of domains locate them in the known three-dimensional structure of a protein, the technique presented here also permits the prediction of domains in proteins of unknown spatial structure, as the construction of the ADM for a given protein requires knowledge of only its amino acid sequence.     

Frequencies of hydrophobic and hydrophilic runs and alternations in proteins of known structure

Patterns of alternation of hydrophobic and polar residues are a profound aspect of amino acid sequences, but a feature not easily interpreted for soluble proteins. Here we report statistics of hydrophobicity patterns in proteins of known structure in a current protein database as compared with results from earlier, more limited structure sets. Previous studies indicated that long hydrophobic runs, common in membrane proteins, are underrepresented in soluble proteins. Long runs of hydrophobic residues remain significantly underrepresented in soluble proteins, with none longer than 16 residues observed. These long runs most commonly occur as buried alpha helices, with extended hydrophobic strands less common. Avoiding aggregation of partially folded intermediates during intracellular folding remains a viable explanation for the rarity of long hydrophobic runs in soluble proteins. Comparison between database editions reveals robustness of statistics on aqueous proteins despite an approximately twofold increase in nonredundant sequences. The expanded database does now allow us to explain several deviations of hydrophobicity statistics from models of random sequence in terms of requirements of specific secondary structure elements. Comparison to prior membrane-bound protein sequences, however, shows significant qualitative changes, with the average hydrophobicity and frequency of long runs of hydrophobic residues noticeably increasing between the database editions. These results suggest that the aqueous proteins of solved structure may represent an essentially complete sample of the universe of aqueous sequences, while the membrane proteins of known structure are not yet representative of the universe of membrane-associated proteins, even by relatively simple measures of hydrophobic patterns.     

Search for structural factors responsible for the stability of proteins from thermophilic organisms

Database including 392 homologous pairs of proteins from thermophilic and mesophilic organisms was created. Using this database we have found that proteins from termophilic organisms contain more atom-atom contacts per residue in comparison with mesophilic homologues. Contribution to increase of the number of contacts gives exterior amino acid residues, accessible for the solvent. Amino acid composition of interior, inaccessible for the solvent, and exterior amino acid residues of proteins from thermophilic and mesophilic organisms were analyzed. We have obtained that exterior residues of proteins from thermophilic organisms contain more such amino acid residues as Lys, Arg and Glu and smaller such amino acid residues as Ala, Asp, Asn. Gln, Ser, and Thr in comparison with proteins from mesophilic organisms. Amino acid compositions of interior residues of considered proteins are not different.     

Primary structure of bovine conglutinin, a member of the C-type animal lectin family

We report the complete amino acid sequence of bovine conglutinin obtained by structural characterization of peptides derived from the protein by various chemical and enzymatic fragmentation methods. The protein consists of 351 amino acid residues including 55 apparent Gly-X-Y repeats with two interruptions. This 171-residue-long collagenous domain separates a short noncollagenous NH2-terminal region of 25 residues from the 155-residue-long globular COOH terminus revealing the structural relation of conglutinin with mannose-binding proteins, pulmonary surfactant-associated proteins, and a complement component C1q. Eight hydroxylysine residues were found in the collagenous domain. All of these hydroxylysine residues which occupy a Y position in a Gly-X-Y triplet are possible glycosylation sites since no phenylthiohydantoin amino acid was identified in automated Edman degradation cycles corresponding to these sites. The noncollagenous COOH domain of conglutinin, on the other hand, contains a carbohydrate recognition domain which shares substantial sequence homology with C-type animal lectins. Conglutinin has the greatest sequence similarity with mannose-binding proteins and pulmonary surfactant-associated proteins.     

Predicting surface exposure of amino acids from protein sequence

The amino acid residues on a protein surface play a key role in interaction with other molecules, determined many physical properties, and constrain the structure of the folded protein. A database of monomeric protein crystal structures was used to teach computer-simulated neural networks rules for predicting surface exposure from local sequence. These trained networks are able to correctly predict surface exposure for 72% of residues in a testing set using a binary model, (buried/exposed) and for 54% of residues using a ternary model (buried/intermediate/exposed). In the ternary model, only 11% of the exposed residues are predicted as buried and only 5% of the buried residues are predicted as exposed. Also, since the networks are able to predict exposure with a quantitative confidence estimate, it is possible to assign exposure for over half of the residues in a binary model with greater than 80% accuracy. Even more accurate predictions are obtained by making a consensus prediction of exposure for a homologous family. The effect of the local environment of an amino acid on its accessibility, though smaller than expected, is significant and accounts for the higher success rate of prediction than obtained with previously used criteria. In the absence of a three-dimensional structure, the ability to predict surface accessibility of amino acids directly from the sequence is a valuable tool in choosing sites of chemical modification or specific mutations and in studies of molecular interaction.     

A simple statistical method for estimating type-II (cluster-specific) functional divergence of protein sequences

Predicting functional amino acid residues in silico is important for comparative genomics. In this paper, we focus on the issue of how to statistically identify cluster-specific amino acid residues that are related to the functional divergence after gene duplication. We approach this problem using a framework based on site-specific shift of amino acid property (type-II functional divergence), as opposed to site-specific shift of evolutionary rate (type-I functional divergence). An efficient statistical procedure is implemented to facilitate the development of phylogenomic database for cluster-specific residues of large-scale protein families. Our method has the following features: 1) statistical testing of the type-II functional divergence and 2) the site-specific Bayesian profile to measure how amino acid residues contribute to type-II (cluster-specific) functional divergence. Consequently, one may obtain the posterior probability for "functional" cluster-specific residues. Case studies are presented and indicate that radical cluster-specific residues are responsible for most of inferred type-II functional divergence, whereas conserved cluster-specific residues appear less than even those imperfect radical cluster-specific residues to this type of functional divergence.     

Analysis of side-chain orientations in homologous proteins

The side-chain conformations of topologically equivalent residues in seven pairs of proteins ranging in sequence homology from 16% to 60% are compared. Both identical and mutated residues are included. For proteins with greater than 40% homology, it is found that at least 80% of the side-chain orientations of identical residues and 75% or more of the mutated residues in each pair of proteins have matching gamma atom dihedral angles (+/- 40 degrees); the comparison is not based strictly on chi 1 angles. Further, if a match is obtained at the gamma position, there is a high probability of matching for the delta atom(s) of the side-chain. For proteins with less than 25% homology the percentages are somewhat lower. Trends observed for conservative substitutions are essentially the same as those noted for mutated residues in general. Side-chain accessibility does not affect the probability of matches of identical residues; however, less accessible pairs of mutated residues have 10 to 20% higher matching probabilities than do exposed residues. Mismatches can frequently be related to large B-factors, certain types of amino acid substitutions, or the appearance of multiple minima on the side-chain potential energy surfaces and are most likely to occur for certain small residues (Ser, Thr, Val). Analysis of all the results makes possible the formulation of a set of rules for side-chain positioning in the modeling of homologous proteins.     

A graphic representation of protein sequence and predicting the subcellular locations of prokaryotic proteins

Zp curve, a three-dimensional space curve representation of protein primary sequence based on the hydrophobicity and charged properties of amino acid residues along the primary sequence is suggested. Relying on the Zp parameters extracted from the three components of the Zp curve and the Bayes discriminant algorithm, the subcellular locations of prokaryotic proteins were predicted. Consequently, an accuracy of 81.5% in the cross-validation test has been achieved using 13 parameters extracted from the curve for the database of 997 prokaryotic proteins. The result is slightly better than that of using the neural network method (80.9%) based on the amino acid composition for the same database. By jointing the amino acid composition and the Zp parameters, the overall predictive accuracy 89.6% can be achieved. It is about 3% higher than that of the Bayes discriminant algorithm based merely on the amino acid composition for the same database. The prediction is also performed with a larger dataset derived from the version 39 SWISS-PROT databank and two datasets with different sequence similarity. Even for the dataset of non-sequence similarity, the improvement can be of 4.4% in the cross-validation test. The results indicate that the Zp parameters are effective in representing the information within a protein primary sequence. The method of extracting information from the primary structure may be useful for other areas of protein studies.