Prediction of protein secondary structure and active sites using the alignment of homologous sequences

The prediction of protein secondary structure (alpha-helices, beta-sheets and coil) is improved by 9% to 66% using the information available from a family of homologous sequences. The approach is based both on averaging the Garnier et al. (1978) secondary structure propensities for aligned residues and on the observation that insertions and high sequence variability tend to occur in loop regions between secondary structures. Accordingly, an algorithm first aligns a family of sequences and a value for the extent of sequence conservation at each position is obtained. This value modifies a Garnier et al. prediction on the averaged sequence to yield the improved prediction. In addition, from the sequence conservation and the predicted secondary structure, many active site regions of enzymes can be located (26 out of 43) with limited over-prediction (8 extra). The entire algorithm is fully automatic and is applicable to all structural classes of globular proteins.     

Refolding of Myoglobins of Nonhomologous Sequences by the Island Model

The folding process of sea hare myoglobin was simulated by the island model, which does not rely on sequence homologies or statistical inference from database of known structure. Sea hare myoglobin has low sequence homology (28%), but high structural similarity, with sperm whale myoglobin, which was already simulated by the island model. Their structural similarity is shown physiochemically from the distribution of hydrophobic-residue pairs, that is, the key pairs for packing of the secondary structures. Irrelevant to the sequence homology, the secondary structures can be packed into the tertiary structure through the hydrophobic interactions among the amino acid pairs responsible for the local structure formation. The results on the two species of myoglobins indicate that, in contrast to other prediction methods, the island model is applicable to any type of protein without extra information other than the distribution of hydrophobic-residue pairs and the positions of the secondary structures. Consequently the present results provide another verification of the validity of the island model for elucidating the mechanisms of protein folding and predicting protein structures.     

Extracting information on folding from the amino acid sequence: consensus regions with preferred conformation in homologous proteins.

It is investigated whether protein segments predicted to have a well-defined conformational preference in the absence of tertiary interactions are conserved in families of homologous proteins. The prediction method follows the procedures of Rooman, M., Kocher, J.-P., and Wodak, S. (preceding paper in this issue). It uses a knowledge-based force field that incorporates only local interactions along the sequence and identifies segments whose lowest energy structure displays a sizable energy gap relative to other computed conformations. In 13 of the protein families and subfamilies considered that are sufficiently homologous to have similar 3D structures, at least one region is consistently predicted as having the same preferred conformation in virtually all family members. These regions are between 4 and 26 residues long. They are often located at chain ends and correspond primarily to segments of secondary structure heavily involved in interactions with the rest of the protein, suggesting that they could act as nuclei around which other parts of the structure would assemble. Experimental data on early folding intermediates or on protein fragments with appreciable structure in aqueous solution are available for more than half of the protein families. Comparison of our results with these data is quite favorable. They reveal that each of the experimentally identified early formed, or independently stable, substructures harbors at least one of the segments consistently predicted as having a preferred conformation by our procedure. The implications of our findings for the conservation of folding pathways in homologous proteins are discussed.     

Conformational preference functions for predicting helices in membrane proteins

A suite of FORTRAN programs, PREF, is described for calculating preference functions from the data base of known protein structures and for comparing smoothed profiles of sequence-dependent preferences in proteins of unknown structure. Amino acid preferences for a secondary structure are considered as functions of a sequence environment. Sequence environment of amino acid residue in a protein is defined as an average over some physical, chemical, or statistical property of its primary structure neighbors. The frequency distribution of sequence environments in the data base of soluble protein structures is approximately normal for each amino acid type of known secondary conformation. An analytical expression for the dependence of preferences on sequence environment is obtained after each frequency distribution is replaced by corresponding Gaussian function. The preference for the α-helical conformation increases for each amino acid type with the increase of sequence environment of buried solvent-accessible surface areas. We show that a set of preference functions based on buried surface area is useful for predicting folding motifs in α-class proteins and in integral membrane proteins. The prediction accuracy for helical residues is 79% for 5 integral membrane proteins and 74% for 11 α-class soluble proteins. Most residues found in transmembrane segments of membrane proteins with known α-helical structure are predicted to be indeed in the helical conformation because of very high middle helix preferences. Both extramembrane and transmembrane helices in the photosynthetic reaction center M and L subunits are correctly predicted. We point out in the discussion that our method of conformational preference functions can identify what physical properties of the amino acids are important in the formation of particular secondary structure elements. © 1993 John Wiley & Sons, Inc.     

Database of homology-derived protein structures and the structural meaning of sequence alignment

The database of known protein three-dimensional structures can be significantly increased by the use of sequence homology, based on the following observations. (1) The database of known sequences, currently at more than 12,000 proteins, is two orders of magnitude larger than the database of known structures. (2) The currently most powerful method of predicting protein structures is model building by homology. (3) Structural homology can be inferred from the level of sequence similarity. (4) The threshold of sequence similarity sufficient for structural homology depends strongly on the length of the alignment. Here, we first quantify the relation between sequence similarity, structure similarity, and alignment length by an exhaustive survey of alignments between proteins of known structure and report a homology threshold curve as a function of alignment length. We then produce a database of homology-derived secondary structure of proteins (HSSP) by aligning to each protein of known structure all sequences deemed homologous on the basis of the threshold curve. For each known protein structure, the derived database contains the aligned sequences, secondary structure, sequence variability, and sequence profile. Tertiary structures of the aligned sequences are implied, but not modeled explicitly. The database effectively increases the number of known protein structures by a factor of five to more than 1800. The results may be useful in assessing the structural significance of matches in sequence database searches, in deriving preferences and patterns for structure prediction, in elucidating the structural role of conserved residues, and in modeling three-dimensional detail by homology.     

Evaluation and improvements in the automatic alignment of protein sequences

The accuracy of protein sequence alignment obtained by applying a commonly used global sequence comparison algorithm is assessed. Alignments based on the superposition of the three-dimensional structures are used as a standard for testing the automatic, sequence-based methods. Alignments obtained from the global comparison of five pairs of homologous protein sequences studied gave 54% agreement overall for residues in secondary structures. The inclusion of information about the secondary structure of one of the proteins in order to limit the number of gaps inserted in regions of secondary structure, improved this figure to 68%. A similarity score of greater than six standard deviation units suggests that an alignment which is greater than 75% correct within secondary structural regions can be obtained automatically for the pair of sequences.     

Amino acid sequence homology applied to the prediction of protein secondary structures, and joint prediction with existing methods

The assumption that homologous segments in different proteins may share a similar conformation is applied to the prediction of secondary structures in proteins. Sequences homologous to a target protein are searched, without allowing any gap, and compared against a number of reference proteins of known three-dimensional structure, and then a conformational state (alpha, beta or coil) for each residue of the protein is predicted by looking at the secondary structure of corresponding homologous segments. This prediction is done in a statistical rather than 'deterministic' way, by assigning the most probable conformation state among homologous data to each residue site of a target protein. A test application for 22 sample proteins yields 60% correctness on the average, a better value in comparison with two other existing methods. Joint prediction combining three methods into one is shown to increase the reliability up to 70%, when only the regions identically predicted with the three methods are taken into account. Application of the present method to 10 proteins of unknown structure is demonstrated.     

Evolved cellular automata for protein secondary structure prediction imitate the determinants for folding observed in nature

We demonstrate the first application of cellular automata to the secondary structure predictions of proteins. Cellular automata use localized interactions to simulate global phenomena, which resembles the protein folding problem where individual residues interact locally to define the global protein conformation. The protein's amino acid sequence was input into the cellular automaton and rules for updating states were evolved using a genetic algorithm. An optimized accuracy (Q3) for the RS126 and CB513 dataset of 58.21% and 56.51%, respectively, could be obtained. Thus, the current work demonstrates the applicability of a rather simple algorithm on a problem as complex as protein secondary structure prediction.     

Combining evolutionary information and neural networks to predict protein secondary structure

Using evolutionary information contained in multiple sequence alignments as input to neural networks, secondary structure can be predicted at significantly increased accuracy. Here, we extend our previous three-level system of neural networks by using additional input information derived from multiple alignments. Using a position-specific conservation weight as part of the input increases performance. Using the number of insertions and deletions reduces the tendency for overprediction and increases overall accuracy. Addition of the global amino acid content yields a further improvement, mainly in predicting structural class. The final network system has a sustained overall accuracy of 71.6% in a multiple cross-validation test on 126 unique protein chains. A test on a new set of 124 recently solved protein structures that have no significant sequence similarity to the learning set confirms the high level of accuracy. The average cross-validated accuracy for all 250 sequence-unique chains is above 72%. Using various data sets, the method is compared to alternative prediction methods, some of which also use multiple alignments: the performance advantage of the network system is at least 6 percentage points in three-state accuracy. In addition, the network estimates secondary structure content from multiple sequence alignments about as well as circular dichroism spectroscopy on a single protein and classifies 75% of the 250 proteins correctly into one of four protein structural classes. Of particular practical importance is the definition of a position-specific reliability index. For 40% of all residues the method has a sustained three-state accuracy of 88%, as high as the overall average for homology modelling. A further strength of the method is greatly increased accuracy in predicting the placement of secondary structure segments. © 1994 Wiley-Liss, Inc.     

Role of evolutionary information in prediction of aromatic-backbone NH interactions in proteins

In this study, an attempt has been made to develop a neural network-based method for predicting segments in proteins containing aromatic-backbone NH (Ar-NH) interactions using multiple sequence alignment. We have analyzed 3121 segments seven residues long containing Ar-NH interactions, extracted from 2298 non-redundant protein structures where no two proteins have more than 25% sequence identity. Two consecutive feed-forward neural networks with a single hidden layer have been trained with standard back-propagation as learning algorithm. The performance of the method improves from 0.12 to 0.15 in terms of Matthews correlation coefficient (MCC) value when evolutionary information (multiple alignment obtained from PSI-BLAST) is used as input instead of a single sequence. The performance of the method further improves from MCC 0.15 to 0.20 when secondary structure information predicted by PSIPRED is incorporated in the prediction. The final network yields an overall prediction accuracy of 70.1% and an MCC of 0.20 when tested by five-fold cross-validation. Overall the performance is 15.2% higher than the random prediction. The method consists of two neural networks: (i) a sequence-to-structure network which predicts the aromatic residues involved in Ar-NH interaction from multiple alignment of protein sequences and (ii) a structure-to structure network where the input consists of the output obtained from the first network and predicted secondary structure. Further, the actual position of the donor residue within the 'potential' predicted fragment has been predicted using a separate sequence-to-structure neural network. Based on the present study, a server Ar_NHPred has been developed which predicts Ar-NH interaction in a given amino acid sequence. The web server Ar_NHPred is available at and (mirror site).     

Use of secondary structural information and C<Emphasis Type="Italic">α</Emphasis>-C<Emphasis Type="Italic">α</Emphasis> distance restraints to model protein structures with MODELLER

Protein secondary structure predictions and amino acid long range contact map predictions from primary sequence of proteins have been explored to aid in modelling protein tertiary structures. In order to evaluate the usefulness of secondary structure and 3D-residue contact prediction methods to model protein structures we have used the known Q3 (alpha-helix, beta-strands and irregular turns/loops) secondary structure information, along with residue-residue contact information as restraints for MODELLER. We present here results of our modelling studies on 30 best resolved single domain protein structures of varied lengths. The results shows that it is very difficult to obtain useful models even with 100% accurate secondary structure predictions and accurate residue contact predictions for up to 30% of residues in a sequence. The best models that we obtained for proteins of lengths 37, 70, 118, 136 and 193 amino acid residues are of RMSDs 4.17, 5.27, 9.12, 7.89 and 9.69, respectively. The results show that one can obtain better models for the proteins which have high percent of alpha-helix content. This analysis further shows that MODELLER restrain optimization program can be useful only if we have truly homologous structure(s) as a template where it derives numerous restraints, almost identical to the templates used. This analysis also clearly indicates that even if we satisfy several true residue-residue contact distances, up to 30% of their sequence length with fully known secondary structural information, we end up predicting model structures much distant from their corresponding native structures.     

The protein structure code: what is its present status?

Current methods of prediction of protein conformation are reviewedand the algorithms on which they rely are presented. For non-homologousproteins and after cross-validation the reported methods exhibita probability index, i.e. the per cent of correctly predictedresidues per predicted residues, of 63–65% with a standarddeviation of the order of 7% for three conformational states—helix,ß-strand and coil. This present limitation in theaccuracy of predictions that use only the information of thelocal sequence can be related essentially to the effect of long-rangeinteractions specific for each protein family. The methods basedon sequence similarity can improve the accuracy of predictionby expressing explicitly the homology of the protein to be predictedwith proteins in the database. In these circumstances the probabilityindex can reach 87% with a standard deviation of 6.6%. Thisproperty can be used for modeling homologous proteins by aidingin amino acid sequence alignments. The prediction of the tertiarystructure of a protein is still limited to the case of modelinga structure based on the known three-dimensional structure ofa homologous protein.     

Use of secondary structural information and C alpha-C alpha distance restraints to model protein structures with MODELLER

Protein secondary structure predictions and amino acid long range contact map predictions from primary sequence of proteins have been explored to aid in modelling protein tertiary structures. In order to evaluate the usefulness of secondary structure and 3D-residue contact prediction methods to model protein structures we have used the known Q3 (alpha-helix,beta-strands and irregular turns/loops) secondary structure information, along with residue-residue contact information as restraints for MODELLER. We present here results of our modelling studies on 30 best resolved single domain protein structures of varied lengths. The results shows that it is very difficult to obtain useful models even with 100% accurate secondary structure predictions and accurate residue contact predictions for up to 30% of residues in a sequence. The best models that we obtained for proteins of lengths 37, 70, 118, 136 and 193 amino acid residues are of RMSDs 4.17, 5.27, 9.12, 7.89 and 9.69,respectively. The results show that one can obtain better models for the proteins which have high percent of alpha-helix content. This analysis further shows that MODELLER restrain optimization program can be useful only if we have truly homologous structure(s) as a template where it derives numerous restraints, almost identical to the templates used. This analysis also clearly indicates that even if we satisfy several true residue-residue contact distances, up to 30%of their sequence length with fully known secondary structural information, we end up predicting model structures much distant from their corresponding native structures.     

Improvement in prediction of solvent accessibility by probability profiles

The capability of predicting folding and conformation of a protein from its primary structure is probably one of the main goals of modern biology. An accurate prediction of solvent accessibility is an intermediate step along this way. A new method for predicting solvent accessibility from single sequence and multiple alignment data is described. The method is based on probability profiles calculated on an amino acid sequence centred on the residue whose accessibility has to be predicted. A profile is constructed for each exposure category considered so as to calculate the probability of a sequence being generated by the different profiles. Prediction accuracy was tested on a variety of protein sets with two- and three-state models. Different thresholds were used according to those adopted by the authors proposing the data sets. The prediction accuracy is significantly improved over existing methods.     

Combining the GOR V algorithm with evolutionary information for protein secondary structure prediction from amino acid sequence

We have modified and improved the GOR algorithm for the protein secondary structure prediction by using the evolutionary information provided by multiple sequence alignments, adding triplet statistics, and optimizing various parameters. We have expanded the database used to include the 513 non-redundant domains collected recently by Cuff and Barton (Proteins 1999;34:508-519; Proteins 2000;40:502-511). We have introduced a variable size window that allowed us to include sequences as short as 20-30 residues. A significant improvement over the previous versions of GOR algorithm was obtained by combining the PSI-BLAST multiple sequence alignments with the GOR method. The new algorithm will form the basis for the future GOR V release on an online prediction server. The average accuracy of the prediction of secondary structure with multiple sequence alignment and full jack-knife procedure was 73.5%. The accuracy of the prediction increases to 74.2% by limiting the prediction to 375 (of 513) sequences having at least 50 PSI-BLAST alignments. The average accuracy of the prediction of the new improved program without using multiple sequence alignments was 67.5%. This is approximately a 3% improvement over the preceding GOR IV algorithm (Garnier J, Gibrat JF, Robson B. Methods Enzymol 1996;266:540-553; Kloczkowski A, Ting K-L, Jernigan RL, Garnier J. Polymer 2002;43:441-449). We have discussed alternatives to the segment overlap (Sov) coefficient proposed by Zemla et al. (Proteins 1999;34:220-223).     

Clustering of amino acids for protein secondary structure prediction

Simple hidden Markov models are proposed for predicting secondary structure of a protein from its amino acid sequence. Since the length of protein conformation segments varies in a narrow range, we ignore the duration effect of length distribution, and focus on inclusion of short range correlations of residues and of conformation states in the models. Conformation-independent and -dependent amino acid coarse-graining schemes are designed for the models by means of proper mutual information. We compare models of different level of complexity, and establish a practical model with a high prediction accuracy.     

Predictions of protein segments with the same aminoacid sequence and different secondary structure: a benchmark for predictive methods

The most stringent test for predictive methods of protein secondary structure is whether identical short sequences that are known to be present with different conformations in different proteins known at atomic resolution can be correctly discriminated. In this study, we show that the prediction efficiency of this type of segments in unrelated proteins reaches an average accuracy per residue ranging from about 72 to 75% (depending on the alignment method used to generate the input sequence profile) only when methods of the third generation are used. A comparison of different methods based on segment statistics (2nd generation methods) and/or including also evolutionary information (3rd generation methods) indicate that the discrimination of the different conformations of identical segments is dependent on the method used for the prediction. Accuracy is similar when methods similarly performing on the secondary structure prediction are tested. When evolutionary information is taken into account as compared to single sequence input, the number of correctly discriminated pairs is increased twofold. The results also highlight the predictive capability of neural networks for identical segments whose conformation differs in different proteins.     

A detailed consideration of a principal domain of vertebrate fibrinogen and its relatives.

Vertebrate fibrinogen is a complex multidomained protein, the structure of which has been inferred mainly from electron microscopy and amino acid sequence studies. Among its most prominent features are two terminal globules, moieties that are mostly composed of the carboxyl-terminal two-thirds of the beta and gamma chains. Sequences homologous to the latter segments are found in several other animal proteins, always as the carboxyl-terminal contributions. An alignment of 15 amino acid sequences from various fibrinogens and related proteins has been used to make judgments about secondary structure. The nature of amino acids at each position in the alignment was used to distinguish alpha helices and beta structure on the one hand from loops and turns on the other, and the resulting assignments compared with predictions of secondary structure by other methods. Additionally, constraints imposed by the locations of cystines, carbohydrate attachment residues, and proteinase-sensitive points provided further insights into the general organization of the postulated secondary structures. Other ancillary data, including the effects of bound calcium and the locations of labeled or variant residues, were also considered. An intriguing similarity to a portion of the recently reported structure of a calcium-dependent lectin is noted.     

A segment alignment approach to protein comparison

MOTIVATION: Local structure segments (LSSs) are small structural units shared by unrelated proteins. They are extensively used in protein structure comparison, and predicted LSSs (PLSSs) are used very successfully in ab initio folding simulations. However, predicted or real LSSs are rarely exploited by protein sequence comparison programs that are based on position-by-position alignments. RESULTS: We developed a SEgment Alignment algorithm (SEA) to compare proteins described as a collection of predicted local structure segments (PLSSs), which is equivalent to an unweighted graph (network). Any specific structure, real or predicted corresponds to a specific path in this network. SEA then uses a network matching approach to find two most similar paths in networks representing two proteins. SEA explores the uncertainty and diversity of predicted local structure information to search for a globally optimal solution. It simultaneously solves two related problems: the alignment of two proteins and the local structure prediction for each of them. On a benchmark of protein pairs with low sequence similarity, we show that application of the SEA algorithm improves alignment quality as compared to FFAS profile-profile alignment, and in some cases SEA alignments can match the structural alignments, a feat previously impossible for any sequence based alignment methods.     

Improvements in a secondary structure prediction method based on a search for local sequence homologies and its use as a model building tool

This report describes an optimised version of a secondary structure prediction method based on local homologies, using a new data base. A 63% prediction accuracy, for three states, was obtained after elimination of the protein to be predicted and all proteins with a percentage identity greater than 22% from the data base. This corresponds to a 5% increase in accuracy on the original method (Levin et al. FEBS Lett. 205 (1986) 303-308). The flexibility of the method to the incorporation of information extraneous to the prediction was demonstrated by the prediction of the homologous proteins in the data base. Using the percentage identity with the protein to be predicted, to weight the relative importance of each protein, for all proteins with a percentage identity greater than 30%, the mean correct prediction per chain was 87%. As a result this algorithm can be used during the molecular modelling process, both to give an idea of the structural similarity between two proteins and as an aid in the determination of the best alignment. Incorporation of the result of a protein folding type assignment based on the global amino-acid composition increased the overall prediction to 66%.