Charalambos Chrysostomou

Fuzzy rules for describing subgroups from Influenza A virus using a multi-objective evolutionary algorithm

Abstract Extraction of biologically-meaningful knowledge is one of the important and challenging tasks in bioinformatics, in particular computational analysis of DNA and protein sequences, in order to identify biological function(s) and behaviour(s) of newly-extracted sequences. Computational intelligence techniques in corporation with sequence-driven features have been applied to tackle the problem and help classify different functional classes of the sequences. In order to study this problem, subgroup discovery algorithms together with a signal processing-based feature extraction method are applied, where the sequences are represented as a signal. The applicability of this method has been studied through four different Neuraminidase genes of Influenza A subtypes, H1N1, H2N2, H3N2 and H5N1. The results yielded not only higher predictive accuracy over these four classes of the proteins but also interpretable rule-based representation of the descriptive model with a significantly reduced feature set driven by means of the signal processing method. Subgroup discovery technique based on evolutionary fuzzy systems is expected to open new areas of research in bioinformatics and further help identify and understand more focused therapeutic protein targets.

Complex Resonant Recognition Model in analysing Influenza a virus subtype protein sequences

Resonant Recognition Method that uses discrete Fourier transform (DFT) and Electron-ion interaction potential amino acid scale (EIIP) is one of the techniques widely used for the analysis of protein sequences. However, DFT that generates complex output (imaginary and real frequency spectra) has shown to produce complementary information in other areas (e.g., ultrasound) were not taken into consideration. Therefore, for the first time, this study is concerned with the development of complex resonant recognition method (CRRM) for the analysis of groups of proteins using their sequence information. As a case study, the method developed is applied to extract characteristic frequency peaks of Influenza A subtypes Neuraminidase gene, for which Influenza A virus subtypes H1N1, H2N2, H3N2 and H5N1 proteins were extracted from Influenza Virus Resource database. The relationships of Influenza A subtypes that appear in CRRM real and imaginary spectra are found to be consistent to the biological link whereas this was not observed in the traditional RRM. H3N2 inherited NA gene from H2N2 and they are found to share the same characteristic frequency as seen in the real spectrum. In addition, H1N1 supplied the NA gene to H5N1 and they also have the same characteristic frequency in the imaginary spectrum. The results clearly show that imaginary part of the CRRM clearly identified similarities and differences between the influenza sub-types at the proteomic level where real part and absolute value of the DFT were incapable of doing so. The results obtained for this study therefore suggest that the CRRM cannot only produce additional biological information but also helps better distinguish biological differences between the families of the proteins. This is hence expected to help better understand mechanisms of the diseases and aid drug/vaccine development.

CISAPS: Complex Informational Spectrum for the Analysis of Protein Sequences

Complex informational spectrum analysis for protein sequences (CISAPS) and its web-based server are developed and presented. As recent studies show, only the use of the absolute spectrum in the analysis of protein sequences using the informational spectrum analysis is proven to be insufficient. Therefore, CISAPS is developed to consider and provide results in three forms including absolute, real, and imaginary spectrum. Biologically related features to the analysis of influenza A subtypes as presented as a case study in this study can also appear individually either in the real or imaginary spectrum. As the results presented, protein classes can present similarities or differences according to the features extracted from CISAPS web server. These associations are probable to be related with the protein feature that the specific amino acid index represents. In addition, various technical issues such as zero-padding and windowing that may affect the analysis are also addressed. CISAPS uses an expanded list of 611 unique amino acid indices where each one represents a different property to perform the analysis. This web-based server enables researchers with little knowledge of signal processing methods to apply and include complex informational spectrum analysis to their work.

Effects of windowing and zero-padding on Complex Resonant Recognition Model for protein sequence analysis

Signal processing techniques such as Fourier Transform have widely been studied and successfully applied in many different areas. Techniques such as zero-padding and windowing have been developed and found very useful to improve the outcome of the signal processing methods. Resonant Recognition Model (RRM) and Complex Resonant Recognition Model (CRRM) that are based on the discrete Fourier Transform and widely used for the analysis of protein sequences do not consider such methods, which can however improve or alter the features extracted from the protein sequences. Therefore, in this paper, an extensive analysis was carried out to investigate into the influence of the zero-padding and windowing on the features extracted from the Complex Resonant Recognition Model. In order to present such effects, five different classes of influenza A virus Neuraminidase genes, which include H1N1, H1N2, H2N2, H3N2 and H5N1 genes, were used as a case study. For each of the Influenza A subtypes, two sets of Common Frequency Peaks (CFP) were extracted, one where windowing is applied and the other one where windowing is suppressed, for each signal length set for the analysis. In order to make all the signals (protein sequence) the same length, zero-padding was used. The signal lengths used in this study are set to 470, which is the maximum protein length, and also 512, 1024, 2048, 4096, 8192 and 16384 for further analysis. The results suggest that the windowing and zero-padding have key impact on CFP extracted from the Influenza A subtypes as the best match with CFP extracted from influenza A subtypes using CRRM is when the signal length of 4096 and windowing were both applied. Therefore, the outcome of this study should be taken into consideration for more accurate and reliable analysis of the protein sequences.

Prediction of protein allergenicity based on signal-processing bioinformatics approach

Current bioinformatics tools accomplish high accuracies in classifying allergenic protein sequences with high homology and generally perform poorly with low homology protein sequences. Although some homologous regions explained Immunoglobulin E (IgE) cross-reactivity in groups of allergens, no universal molecular structure could be associated with allergenicity. In addition, studies have showed that cross-reactivity is not directly linked to the homology between protein sequences. Therefore, a new homology independent method needs to be developed to determine if a protein is an allergen or not. The aim of this study is therefore to differentiate sets of allergenic and non-allergenic proteins using a signal-processing based bioinformatics approach. In this paper, a new method was proposed for characterisation and classification of allergenic protein sequences. For this method hydrophobicity amino acid index was used to encode proteins to numerical sequences and Discrete Fourier Transform to extract features for each protein. Finally, a classifier was constructed based on Support Vector Machines. In order to demonstrate the applicability of the proposed method 857 allergen and 1000 non-allergen proteins were collected from UniProt online database. The results obtained from the proposed method yielded: MCC: 0.752 ± 0.007, Specificity: 0.912 ± 0.005, Sensitivity: 0.835 ± 0.008 and Total Accuracy: 87.65% ± 0.004.

Construction of protein dendrograms based on amino acid indices and Discrete Fourier Transform.

From the literature, existing methods use pairwise percent identity to identify the percentage of similarity between two protein sequences, in order to create a dendrogram. As this is a parametric method of measuring the similarities between proteins, and different parameter may yield different results, this method does not guarantee that the global optimal similarity values will be found. As protein dendrogram construction is used in other areas, such as multiple protein sequence alignments, it is very important that the most related protein sequences to be identified and align first. Furthermore, by using the pairwise percent identity of the protein sequences to construct the dendrograms, the physical characteristics of protein sequences and amino acids are not considered. In this paper, a new method was proposed for constructing protein sequence dendrograms. For this method, Discrete Fourier Transform, was used to construct the distance matrix in combination with the multiple amino acid indices that were used to encode protein sequences into numerical sequences. In order to show the applicability and robustness of the proposed method, a case study was presented by using nine Cluster of Differentiation 4 protein sequences extracted from the UniProt online database.

Construction of protein distance matrix based on amino acid indices and Discrete Fourier Transform

Protein distance matrix is widely used in various protein sequence analyses, and mainly obtained by using pairwise sequence alignment scores or protein sequence homology, which fail to take into consideration of individual physical characteristics of protein sequences and amino acids, or a combination of these features. In this paper, a new method is therefore proposed for constructing protein distance matrix based on natural amino acid indices in combination with Discrete Fourier Transform (DFT). For the proposed method, protein distance matrices can be generated using any given set of amino acid indices, each one of which represents a unique biological feature of protein sequences. In this study, the results are based on the combination of 25 widely accepted amino acid indices, which produced the best results, according to the biological relationships between proteins. As a case study 26 Cluster of Differentiation 4 (CD4) protein sequences were used in order to construct a distance matrix based on the proposed method. The results show that the pairwise relationship between CD4 protein sequences remain the same in comparison with their pairwise percent identity. For another group of protein sequences the pairwise relationship between CD4 protein sequences dramatically changed with the proposed method in comparison to the pairwise percent identity. The proposed distance matrix has been shown to have a positive impact on these case studies and therefore is expected to be useful in several fields such as multiple protein sequence alignment and phylogenetic analysis, where an accurate distance matrix based on natural generalized protein properties plays an important role.

Surface Reconstruction Techniques Using Neural Networks to Recover Noisy 3D Scenes

This paper presents a novel neural network approach to recovering of 3D surfaces from single gray scale images. The proposed neural network uses photometric stereo to estimate local surfaces orientation for surfaces at each point of the surface that was observed from same viewpoint but with different illumination direction in surfaces that follow the Lambertian reflection model. The parameters for the neural network are a 3x3 brightness patch with pixel values of the image and the light source direction. The light source direction of the surface is calculated using two different approaches. The first approach uses a mathematical method and the second one a neural network method. The images used to test the neural network were both synthetic and real images. Only synthetic images were used to compare the approaches mainly because the surface was known and the error could be calculated. The results show that the proposed neural network is able to recover the surface with a highly accurately estimate.

Extension to Distributed Annotation System: Summary command

Using the current version of the Distributed Annotation System (DAS) protocol to obtain data from large regions of interest from remote DAS servers can be a time and resource consuming process. Therefore it would be useful to know the amounts and types of features that exist in a region of interest before the DAS request is made. In the current DAS protocol (1.6), the types of data that exist with a DAS source can be obtained before the complete set of features are requested for a specific region using the feature command. Depending on the implementation of the types command in the DAS server, the number of features across the segment can also be obtained. However, counting of features is computationally intensive for every user request and so most DAS servers do not include it in their implementation. For these DAS servers in order for the user to obtain counts of features, the complete set of features needs to be obtained and re-analysed. Additionally in the current DAS protocol no parameter exists to include or exclude the count of features per type when needed. In this paper, an addition to DAS protocol is proposed and implemented in order to request a summary of the features that exist within a region of interest when needed. The summary command was implemented in order to broaden the functionalities, and extend the flexibility of the current DAS protocol. The summary functionality can conserve time and resources, especially for large regions of interest.

Signal-processing-based bioinformatics approach for the identification of influenza A virus subtypes in Neuraminidase genes

Neuraminidase (NA) genes of influenza A virus is a highly potential candidate for antiviral drug development that can only be realized through true identification of its sub-types. In this paper, in order to accurately detect the sub-types, a hybrid predictive model is therefore developed and tested over proteins obtained from the four subtypes of the influenza A virus, namely, H1N1, H2N2, H3N2 and H5N1 that caused major pandemics in the twentieth century. The predictive model is built by the following four main steps; (i) decoding the protein sequences into numerical signals by means of EIIP amino acid scale, (ii) analysing these signals (protein sequences) by using Discrete Fourier Transform (DFT) and extracting DFT-based features, (iii) selecting more influential sub-set of the features by using the F-score statistical feature selection method, and finally (iv) building a predictive model on the feature sub-set by using support vector machine classifier. The protein sequences were chosen as to be of high percentage identity that they demonstrate within individual influenza subtype classes and high variation that they display in the percentage identity. This makes the proteins very difficult to distinguish from each other even they belong to different subtypes. Given this set of the proteins, the predictive model yielded 98.3% accuracy based on a 5-fold cross validation. This also results in a twenty feature sub-set that can also help reveal spectral characteristics of the subtypes. The proposed model is promising and can easily be generalized for other similar studies.