A feature vector integration approach for a generalized support vector machine pairwise homology algorithm

Due to the exponential growth of sequenced genomes, the need to quickly provide accurate annotation for existing and new sequences is paramount to facilitate biological research. Current sequence comparison approaches fail to detect homologous relationships when sequence similarity is low. Support vector machine (SVM) algorithms approach this problem by transforming all proteins into a feature space of equal dimension based on protein properties, such as sequence similarity scores against a basis set of proteins or motifs. This multivariate representation of the protein space is then used to build a classifier specific to a pre-defined protein family. However, this approach is not well suited to large-scale annotation. We have developed a SVM approach that formulates remote homology as a single classifier that answers the pairwise comparison problem by integrating the two feature vectors for a pair of sequences into a single vector representation that can be used to build a classifier that separates sequence pairs into homologs and non-homologs. This pairwise SVM approach significantly improves the task of remote homology detection on the benchmark dataset, quantified as the area under the receiver operating characteristic curve; 0.97 versus 0.73 and 0.70 for PSI-BLAST and Basic Local Alignment Search Tool (BLAST), respectively.     

Letter to the Editor

A new two-dimensional graphical representation of protein sequences is introduced. Twenty concentric evenly spaced circles divided by n radial lines into equal divisions are selected to represent any protein sequence of length n. Each circle represents one of the different 20 amino acids, and each radial line represents a single amino acid of the protein sequence. An efficient numerical method based on the graph is proposed to measure the similarity between two protein sequences. To prove the accuracy of our approach, the method is applied to NADH dehydrogenase subunit 5 (ND5) proteins of nine different species and 24 transferrin sequences from vertebrates. High values of correlation coefficient between our results and the results of ClustalW are obtained (approximately perfect correlations). These values are higher than the values obtained in many other related works.     

ALIGN_MTX—An optimal pairwise textual sequence alignment program,adapted for using in sequence-structure alignment

The presented program ALIGN_MTX makes alignment of two textual sequences with an opportunity to use any several characters for the designation of sequence elements and arbitrary user substitution matrices. It can be used not only for the alignment of amino acid and nucleotide sequences but also for sequence-structure alignment used in threading, amino acid sequence alignment, using preliminary known PSSM matrix, and in other cases when alignment of biological or non-biological textual sequences is required. This distinguishes it from the majority of similar alignment programs that make, as a rule, alignment only of amino acid or nucleotide sequences represented as a sequence of single alphabetic characters. ALIGN_MTX is presented as downloadable zip archive at http://www.imbbp.org/software/ALIGN_MTX/ and available for free use.As application of using the program, the results of comparison of different types of substitution matrix for alignment quality in distantly related protein pair sets were presented. Threading matrix SORDIS, based on side-chain orientation in relation to hydrophobic core centers with evolutionary change-based substitution matrix BLOSUM and using multiple sequence alignment information position-specific score matrices (PSSM) were taken for test alignment accuracy. The best performance shows PSSM matrix, but in the reduced set with lower sequence similarity threading matrix SORDIS shows the same performance and it was shown that combined potential with SORDIS and PSSM can improve alignment quality in evolutionary distantly related protein pairs.     

Testing homology with Contact Accepted mutatiOn (CAO): a contact-based Markov model of protein evolution

Point Accepted Mutation (PAM) is the Markov model of amino acid replacements in proteins introduced by Dayhoff and her co-workers (Dayhoff et al., 1978). The PAM matrices and other matrices based on the PAM model have been widely accepted as the standard scoring system of protein sequence similarity in protein sequence alignment tools. Here, we present Contact Accepted mutatiOn (CAO), a Markov model of protein residue contact mutations. The CAO model simulates the interchanging of structurally defined side-chain contacts, and introduces additional structural information into protein sequence alignments. Therefore, similarities between structurally conserved sequences can be detected even without apparent sequence similarity. CAO has been benchmarked on the HOMSTRAD database and a subset of the CATH database, by comparing sequence alignments with reference alignments derived from structural superposition. CAO yields scores that reflect coherently the structural quality of sequence alignments, which has implications particularly for homology modelling and threading techniques.     

A novel graphical representation of protein sequences and its application

On the basis of information on the evolution of the 20 amino acids and their physiochemical characteristics, we propose a new two-dimensional (2D) graphical representation of protein sequences in this article. By this representation method, we use 2D data to represent three-dimensional information constructed by the amino acids' evolution index, the class information of amino acid based on physiochemical characteristics, and the order of the amino acids appearing in the protein sequences. Then, using discrete Fourier transform, the sequence signals with different lengths can be transformed to the frequency domain, in which the sequences are with the same length. A new method is used to analyze the protein sequence similarity and to predict the protein structural class. The experiments indicate that our method is effective and useful.     

The graphical representation of protein sequences based on the physicochemical properties and its applications

Based on the chaos game representation, a 2D graphical representation of protein sequences was introduced in which the 20 amino acids are rearranged in a cyclic order according to their physicochemical properties. The Euclidean distances between the corresponding amino acids from the 2‐D graphical representations are computed to find matching (or conserved) fragments of amino acids between the two proteins. Again, the cumulative distance of the 2D‐graphical representations is defined to compare the similarity of protein. And, the examination of the similarity among sequences of the ND5 proteins of nine species shows the utility of our approach. © 2010 Wiley Periodicals, Inc. J Comput Chem, 2010     

Improving protein structural class prediction using novel combined sequence information and predicted secondary structural features

Protein structural class prediction solely from protein sequences is a challenging problem in bioinformatics. Numerous efficient methods have been proposed for protein structural class prediction, but challenges remain. Using novel combined sequence information coupled with predicted secondary structural features (PSSF), we proposed a novel scheme to improve prediction of protein structural classes. Given an amino acid sequence, we first transformed it into a reduced amino acid sequence and calculated its word frequencies and word position features to combine novel sequence information. Then we added the PSSF to the combine sequence information to predict protein structural classes. The proposed method was tested on four benchmark datasets in low homology and achieved the overall prediction accuracies of 83.1%, 87.0%, 94.5%, and 85.2%, respectively. The comparison with existing methods demonstrates that the overall improvements range from 2.3% to 27.5%, which indicates that the proposed method is more efficient, especially for low-homology amino acid sequences.     

An alignment-free method to find similarity among protein sequences via the general form of Chou’s pseudo amino acid composition

In this paper, we propose a method to create the 60-dimensional feature vector for protein sequences via the general form of pseudo amino acid composition. The construction of the feature vector is based on the contents of amino acids, total distance of each amino acid from the first amino acid in the protein sequence and the distribution of 20 amino acids. The obtained cosine distance metric (also called the similarity matrix) is used to construct the phylogenetic tree by the neighbour joining method. In order to show the applicability of our approach, we tested it on three proteins: 1) ND5 protein sequences from nine species, 2) ND6 protein sequences from eight species, and 3) 50 coronavirus spike proteins. The results are in agreement with known history and the output from the multiple sequence alignment program ClustalW, which is widely used. We have also compared our phylogenetic results with six other recently proposed alignment-free methods. These comparisons show that our proposed method gives a more consistent biological relationship than the others. In addition, the time complexity is linear and space required is less as compared with other alignment-free methods that use graphical representation. It should be noted that the multiple sequence alignment method has exponential time complexity.     

An alignment-free measure based on physicochemical properties of amino acids for protein sequence comparison

Sequence comparison is an important topic in bioinformatics. With the exponential increase of biological sequences, the traditional protein sequence comparison methods — the alignment methods become limited, so the alignment-free methods are widely proposed in the past two decades. In this paper, we considered not only the six typical physicochemical properties of amino acids, but also their frequency and positional distribution. A 51-dimensional vector was obtained to describe the protein sequence. We got a pairwise distance matrix by computing the standardized Euclidean distance, and discriminant analysis and phylogenetic analysis can be made. The results on the Influenza A virus and ND5 datasets indicate that our method is accurate and efficient for classifying proteins and inferring the phylogeny of species.     

MOLECULAR CHARACTERIZATION OF RICE Wx GENE

The complete nucleotide (nt) sequence of the rice waxy(Wx) gene, which is responsible for the synthesis of amylose in endosperm and pollen, has been determined by a combination of restriction mapping and nt sequence analysis of two overlapping genomic DNA clones. The entire gene is about 5.5 kb in length. The alignment of the nt sequence of the Wx gene from rice with those of maize (Klsgen, R. B. et al.) and barley (Rohde, W. et al.) revealed the presence of thirteen introns and fourteen exons. The full-length of Wx protein in cluding transit peptide is 609 amino acid (aa) residues. The calculated molecular weight of rice Wx preprotein is about 72 kD. There is no significant difference between the similarity scores of the aa sequence deduced from the rice Wx gene compared with those of maize and barley. However, the nt sequences of the 5'-end upstream, 3'-end downstream and introns of the rice Wx gene, as well as the aa sequence of the transit peptide region of the Wx preprotein have low similarity scor     

A statistical model for predicting protein folding rates from amino acid sequence with structural class information

Prediction of protein folding rates from amino acid sequences is one of the most important challenges in molecular biology. In this work, I have related the protein folding rates with physical-chemical, energetic and conformational properties of amino acid residues. I found that the classification of proteins into different structural classes shows an excellent correlation between amino acid properties and folding rates of two- and three-state proteins, indicating the importance of native state topology in determining the protein folding rates. I have formulated a simple linear regression model for predicting the protein folding rates from amino acid sequences along with structural class information and obtained an excellent agreement between predicted and experimentally observed folding rates of proteins; the correlation coefficients are 0.99, 0.96 and 0.95, respectively, for all-alpha, all-beta and mixed class proteins. This is the first available method, which is capable of predicting the protein folding rates just from the amino acid sequence with the aid of generic amino acid properties and structural class information.     

Comparison of non-sequential sets of protein residues

A methodology for performing sequence-free comparison of functional sites in protein structures is introduced. The method is based on a new notion of similarity among superimposed groups of amino acid residues that evaluates both geometry and physico-chemical properties. The method is specifically designed to handle disconnected and sparsely distributed sets of residues. A genetic algorithm is employed to find the superimposition of protein segments that maximizes their similarity. The method was evaluated by performing an all-to-all comparison on two separate sets of ligand-binding sites, comprising 47 protein-FAD (Flavin-Adenine Dinucleotide) and 64 protein-NAD (Nicotinamide-Adenine Dinucleotide) complexes, and comparing the results with those of an existing sequence-based structural alignment tool (TM-Align). The quality of the two methodologies is judged by the methods’ capacity to, among other, correctly predict the similarities in the protein-ligand contact patterns of each pair of binding sites. The results show that using a sequence-free method significantly improves over the sequence-based one, resulting in 23 significant binding-site homologies being detected by the new method but ignored by the sequence-based one.     

MMM-QSAR recognition of ribonucleases without alignment: comparison with an HMM model and isolation from Schizosaccharomyces pombe, prediction, and experimental assay of a new sequence

The study of type III RNases constitutes an important area in molecular biology. It is known that the pac1+ gene encodes a particular RNase III that shares low amino acid similarity with other genes despite having a double-stranded ribonuclease activity. Bioinformatics methods based on sequence alignment may fail when there is a low amino acidic identity percentage between a query sequence and others with similar functions (remote homologues) or a similar sequence is not recorded in the database. Quantitative structure-activity relationships (QSAR) applied to protein sequences may allow an alignment-independent prediction of protein function. These sequences of QSAR-like methods often use 1D sequence numerical parameters as the input to seek sequence-function relationships. However, previous 2D representation of sequences may uncover useful higher-order information. In the work described here we calculated for the first time the spectral moments of a Markov matrix (MMM) associated with a 2D-HP-map of a protein sequence. We used MMMs values to characterize numerically 81 sequences of type III RNases and 133 proteins of a control group. We subsequently developed one MMM-QSAR and one classic hidden Markov model (HMM) based on the same data. The MMM-QSAR showed a discrimination power of RNAses from other proteins of 97.35% without using alignment, which is a result as good as for the known HMM techniques. We also report for the first time the isolation of a new Pac1 protein (DQ647826) from Schizosaccharomyces pombe strain 428-4-1. The MMM-QSAR model predicts the new RNase III with the same accuracy as other classical alignment methods. Experimental assay of this protein confirms the predicted activity. The present results suggest that MMM-QSAR models may be used for protein function annotation avoiding sequence alignment with the same accuracy of classic HMM models.     

Sequence classification of water channels and related proteins in view of functional predictions

We have worked with a classification method based upon a notion of probabilistic similarity or “likelihood of similarity” between aligned sequences. One important parameter, among others, affecting the sequence similarities and hence the classification results is the amino acid similarity matrix. We present a method for choosing the most adapted matrix to classify protein sequences. This method has been applied to the transmembrane channels of the major intrinsic protein (MIP) family. At present, two functional subgroups have been well characterized in this family: (1) specific water transport by the aquaporins and (2) small neutral solutes transport. The aim of the present study is to show the usefulness of the classification method in the prediction of sequence segments important for substrate selectivity. Moreover, we show that this method can also be used to predict the function of undetermined MIP proteins. The method could be applied to other protein families as well. Received: 24 April 1998 / Accepted: 4 August 1998 / Published online: 2 November 1998     

Computer analysis of automated Edman degradation and amino acid analysis data

Computer programs are described that allow facile analysis of data from a protein sequencer and amino acid analyzer. The sequencer program provides automated sequence interpretation while requiring minimal user interaction. The program serves as a powerful aid in deciphering mixture sequences and allows routine monitoring of sequencer performance. The computer program for amino acid analysis data provides the following calculations: mole percent, protein concentration and residues per mole with comparison between theoretical and calculated values. A plot of molecular weight versus deviation from integer values is calculated providing a measure of peptide or protein purity.     

SVM-based detection of distant protein structural relationships using pairwise probabilistic suffix trees

A new method based on probabilistic suffix trees (PSTs) is defined for pairwise comparison of distantly related protein sequences. The new definition is adopted in a discriminative framework for protein classification using pairwise sequence similarity scores in feature encoding. The framework uses support vector machines (SVMs) to separate structurally similar and dissimilar examples. The new discriminative system, which we call as SVM-PST, has been tested for SCOP family classification task, and compared with existing discriminative methods SVM-BLAST and SVM-Pairwise, which use BLAST similarity scores and dynamic-programming-based alignment scores, respectively. Results have shown that SVM-PST is more accurate than SVM-BLAST and competitive with SVM-Pairwise. In terms of computational efficiency, PST-based comparison is much better than dynamic-programming-based alignment. We also compared our results with the original family-based PST approach from which we were inspired. The present method provides a significantly better solution for protein classification in comparison with the family-based PST model.