Ronesh Sharma

Predict Gram-Positive and Gram-Negative Subcellular Localization via Incorporating Evolutionary Information and Physicochemical Features Into Chou's General PseAAC

In this study, we used structural and evolutionary based features to represent the sequences of gram-positive and gram-negative subcellular localizations. To do this, we proposed a normalization method to construct a normalize Position Specific Scoring Matrix (PSSM) using the information from original PSSM. To investigate the effectiveness of the proposed method we compute feature vectors from normalize PSSM and by applying support vector machine (SVM) and naïve Bayes classifier, respectively, we compared achieved results with the previously reported results. We also computed features from original PSSM and normalized PSSM and compared their results. The archived results show enhancement in gram-positive and gram-negative subcellular localizations. Evaluating localization for each feature, our results indicate that employing SVM and concatenating features (amino acid composition feature, Dubchak feature (physicochemical-based features), normalized PSSM based auto-covariance feature and normalized PSSM based bigram feature) have higher accuracy while employing naïve Bayes classifier with normalized PSSM based auto-covariance feature proves to have high sensitivity for both benchmarks. Our reported results in terms of overall locative accuracy is 84.8% and overall absolute accuracy is 85.16% for gram-positive dataset; and, for gram-negative dataset, overall locative accuracy is 85.4% and overall absolute accuracy is 86.3%.

Decimation Filter with Common Spatial Pattern and Fishers Discriminant Analysis for Motor Imagery Classification

Brain Computer Interface (BCI) system converts thoughts into commands for driving external device with Electroencephalography (EEG). This paper presents the use of decimation filters for filtering the EEG signal. Common Spatial Pattern (CSP) technique is used to transform the filtered signal to a new time series in order to have optimal variance for the discrimination of different tasks. Fishers Discriminant Analysis (FDA) is applied to the CSP features and the FDA scores are fed to a Support Vector Machine (SVM) classifier. The method is evaluated on BCI Competition III Dataset IVa and compared with other related state-of-the-art approaches. The results show that our method outperforms all other approaches in terms of average classification error rate. Compared to best performing method that uses only CSP features, the results obtained in this research offer on average a reduction of 1.07% in the classification error rate.

Predicting MoRFs in protein sequences using HMM profiles

BackgroundIntrinsically Disordered Proteins (IDPs) lack an ordered three-dimensional structure and are enriched in various biological processes. The Molecular Recognition Features (MoRFs) are functional regions within IDPs that undergo a disorder-to-order transition on binding to a partner protein. Identifying MoRFs in IDPs using computational methods is a challenging task.MethodsIn this study, we introduce hidden Markov model (HMM) profiles to accurately identify the location of MoRFs in disordered protein sequences. Using windowing technique, HMM profiles are utilised to extract features from protein sequences and support vector machines (SVM) are used to calculate a propensity score for each residue. Two different SVM kernels with high noise tolerance are evaluated with a varying window size and the scores of the SVM models are combined to generate the final propensity score to predict MoRF residues. The SVM models are designed to extract maximal information between MoRF residues, its neighboring regions (Flanks) and the remainder of the sequence (Others).ResultsTo evaluate the proposed method, its performance was compared to that of other MoRF predictors; MoRFpred and ANCHOR. The results show that the proposed method outperforms these two predictors.ConclusionsUsing HMM profile as a source of feature extraction, the proposed method indicates improvement in predicting MoRFs in disordered protein sequences.

MoRFPred-plus: Computational Identification of MoRFs in Protein Sequences using Physicochemical Properties and HMM profiles

MOTIVATION Intrinsically Disordered Proteins (IDPs) lack stable tertiary structure and they actively participate in performing various biological functions. These IDPs expose short binding regions called Molecular Recognition Features (MoRFs) that permit interaction with structured protein regions. Upon interaction they undergo a disorder-to-order transition as a result of which their functionality arises. Predicting these MoRFs in disordered protein sequences is a challenging task. METHOD In this study, we present MoRFpred-plus, an improved predictor over our previous proposed predictor to identify MoRFs in disordered protein sequences. Two separate independent propensity scores are computed via incorporating physicochemical properties and HMM profiles, these scores are combined to predict final MoRF propensity score for a given residue. The first score reflects the characteristics of a query residue to be part of MoRF region based on the composition and similarity of assumed MoRF and flank regions. The second score reflects the characteristics of a query residue to be part of MoRF region based on the properties of flanks associated around the given residue in the query protein sequence. The propensity scores are processed and common averaging is applied to generate the final prediction score of MoRFpred-plus. RESULTS Performance of the proposed predictor is compared with available MoRF predictors, MoRFchibi, MoRFpred, and ANCHOR. Using previously collected training and test sets used to evaluate the mentioned predictors, the proposed predictor outperforms these predictors and generates lower false positive rate. In addition, MoRFpred-plus is a downloadable predictor, which makes it useful as it can be used as input to other computational tools. AVAILABILITY https://github.com/roneshsharma/MoRFpred-plus/wiki/MoRFpred-plus:-Download.

OPAL: prediction of MoRF regions in intrinsically disordered protein sequences

Motivation Intrinsically disordered proteins lack stable 3-dimensional structure and play a crucial role in performing various biological functions. Key to their biological function are the molecular recognition features (MoRFs) located within long disordered regions. Computationally identifying these MoRFs from disordered protein sequences is a challenging task. In this study, we present a new MoRF predictor, OPAL, to identify MoRFs in disordered protein sequences. OPAL utilizes two independent sources of information computed using different component predictors. The scores are processed and combined using common averaging method. The first score is computed using a component MoRF predictor which utilizes composition and sequence similarity of MoRF and non-MoRF regions to detect MoRFs. The second score is calculated using half-sphere exposure (HSE), solvent accessible surface area (ASA) and backbone angle information of the disordered protein sequence, using information from the amino acid properties of flanks surrounding the MoRFs to distinguish MoRF and non-MoRF residues. Results OPAL is evaluated using test sets that were previously used to evaluate MoRF predictors, MoRFpred, MoRFchibi and MoRFchibi-web. The results demonstrate that OPAL outperforms all the available MoRF predictors and is the most accurate predictor available for MoRF prediction. It is available at http://www.alok-ai-lab.com/tools/opal/. Contact ashwini@hgc.jp or alok.sharma@griffith.edu.au. Supplementary information Supplementary data are available at Bioinformatics online.

Importance of dimensionality reduction in protein fold recognition

Interpreting tertiary structure of a protein has been a crucial task in the field of biosciences. This problem can be addressed by detecting protein folds which is considered as an intermediate step in the tertiary structure prediction. From the perspective of computational sciences, the protein fold recognition can be subdivided in two steps: 1) feature extraction of protein sequences, and 2) identifying extracted features using appropriate classifiers. These steps are important to accurately identify folds of a novel protein sequence. In order to fully characterize a protein sequence, the number of features required is large and sometimes even unmanageable. This high dimensionality of features is difficult to process using conventional classifiers. Therefore, it is a challenge to develop and apply dimensionality reduction techniques for protein fold recognition. In this paper, we have emphasized the importance of dimensionality reduction techniques (DRTs) for protein fold recognition. To narrate, we have compared the recognition performance without DRT and with DRT on 3 benchmark datasets.

OPAL+: Length-Specific MoRF Prediction in Intrinsically Disordered Protein Sequences

Intrinsically disordered proteins (IDPs) contain long unstructured regions, which play an important role in their function. These intrinsically disordered regions (IDRs) participate in binding events through regions called molecular recognition features (MoRFs). Computational prediction of MoRFs helps identify the potentially functional regions in IDRs. In this study, OPAL+, a novel MoRF predictor, is presented. OPAL+ uses separate models to predict MoRFs of varying lengths along with incorporating the hidden Markov model (HMM) profiles and physicochemical properties of MoRFs and their flanking regions. Together, these features help OPAL+ achieve a marginal performance improvement of 0.4–0.7% over its predecessor for diverse MoRF test sets. This performance improvement comes at the expense of increased run time as a result of the requirement of HMM profiles. OPAL+ is available for download at https://github.com/roneshsharma/OPAL-plus/wiki/OPAL-plus-Download.