Archana Chowdhury

A Bacterial Evolutionary Algorithm for automatic data clustering

This paper describes an evolutionary clustering algorithm, which can partition a given dataset automatically into the optimal number of groups through one shot of optimization. The proposed method is based on an evolutionary computing technique known as the Bacterial Evolutionary Algorithm (BEA). The BEA draws inspiration from a biological phenomenon of microbial evolution. Unlike the conventional mutation, crossover and selection operaions in a GA (Genetic Algorithm), BEA incorporates two special operations for evolving its population, namely the bacterial mutation and the gene transfer operation. In the present context, these operations have been modified so as to handle the variable lengths of the chromosomes that encode different cluster groupings. Experiments were done with several synthetic as well as real life data sets including a remote sensing satellite image data. The results estabish the superiority of the proposed approach in terms of final accuracy.

Protein-protein interaction network prediction using stochastic learning automata induced differential evolution

Display Omitted Formulation of PPI problem as a single objective optimization problem.Two novel approaches are proposed to predict a PPI network by independently maximizing individual objectives.Individual objectives are optimized by differential evolution (DE) and stochastic learning automata (SLA).DE is employed to globally explore the search space and SLA for adaptive tuning of the control parameters of the algorithm.The proposed technique outperforms the existing PPI prediction methods. Protein-protein interactions (PPIs) are of biological interest for their active participation in coordinating a number of cellular processes in living organisms. This paper attempts to formulate PPIs as an optimization problem with an aim to independently maximize (a) the stability of a complex formed by two proteins predicted to be interacting, (b) the difference between their individual accessible solvent area and that of the corresponding protein-protein complex, (c) their functional similarity and d) the occurrence of their interacting domain pairs. The novelty of the paper lies in ranking the set of PPI networks, obtained through independently optimizing individual objectives, using two approaches. The first approach is concerned with identifying the equally good PPI networks based on their fitness-based ranks with respect to individual four objectives. The second approach aims at sorting the PPI networks based on their fuzzy memberships to satisfy individual four objectives. The paper also proposes a novel single objective optimization algorithm to optimize individual objectives, influencing the true prediction of a PPI network. The proposed algorithm is realized by an amalgamation of the differential evolution and the stochastic learning automata, where the former is employed to globally explore the search space and the latter for the adaptive tuning of the control parameter of the algorithm. The proposed technique outperforms the existing methods, including relative specific similarity, domain cohesion coupling, random decision forest, fuzzy support vector machine and evolutionary/swarm algorithm based approaches, with respect to both sensitivity and specificity.

Muti-objective evolutionary approach of ligand design for protein-ligand docking problem

The paper addresses a novel approach to protein-ligand docking problem using Non-dominated Sorting Bee Colony optimization algorithm. In this work, protein-ligand docking is formulated as a multi-objective optimization problem. The docking energy, molecular weight and oral bioavailability are used as three scoring functions for the solutions. Results are demonstrated for six different target proteins both numerically and pictorially. Experimental results reveal that the proposed method outperforms Multi-Objective Particle Swarm Optimization, Non-dominated Sorting Genetic Algorithm-II and Artificial Bee Colony based ligand design method considering the three objectives of the evolved molecules.

A modified bat algorithm to predict Protein-Protein Interaction network

This paper provides a novel approach to predict the Protein-Protein Interaction (PPI) network using a modified version of the Bat Algorithm. The attractive trait of the proposed approach is that it attempts to analyze the impact of physicochemical properties, structural features and evolutionary relationship of proteins, to predict the PPI network. Computer simulations reveal that our proposed method effectively predicts the PPI of Saccharomyces Cerevisiae with a sensitivity of (0.85) and specificity of (0.87) and outperforms other state-of-art methodologies.

Protein Function Prediction Using Adaptive Swarm Based Algorithm

The center of attention of the research in bioinformatics has been towards understanding the biological mechanisms and protein functions. Recently high throughput experimental methods have provided many protein-protein interaction networks which need to be analyzed to provide an insight into the functional role of proteins in living organism. One of the important problems of post-genomic era is to predict the functions of unannotated proteins. In this paper we propose a novel approach for protein function prediction by utilizing the fact that most of the proteins which are connected in protein-protein interaction network, tend to have similar functions. The method randomly associates unannotated protein with functions from the possible set of functions. Our approach, Artificial Bee Colony with Temporal Difference Q-Learning (ABC-TDQL), then optimizes the score function which incorporates the extent of similarity between the set of functions of unannotated protein and annotated protein, to associate a function to an unannotated protein. The approach was utilized to predict protein function of Saccharomyces Cerevisiae and the experimental results reveal that our proposed method outperforms other algorithms in terms of precession, recall and F-value.

Differential evolution induced many objective optimization

We propose a novel approach to solve the many objective optimization (MaOO) problem using a ranking policy, instead of the Pareto ranking, supposing that a solution is unlikely to perform well for all objectives in a MaOO problem. A solution is thus evolved with respect to a specific objective only, which it may proficiently optimize. First, all objectives of the MaOO problem are individually optimized by evolutionary algorithms in parallel. The second step is concerned with judiciously selecting and filtering the quality solutions obtained by individual optimization of all objectives in parallel. A unique ranking policy is proposed to grade the members of the union set of quality solutions based on their extent of optimization of individual objectives. The evolutionary algorithm used for parallel optimization of all objectives in a MaOO here has been realized with differential evolution (DE). The mutation strategy of DE is also amended here with an aim to allow controlled communication between population members, concerned with parallel optimization of different objectives of a MaOO problem. Experiments undertaken with DTLZ and WFG test suits reveal that the proposed algorithm outperforms the state-of-art techniques with respect to inverted generational distance and hypervolume metrics.

Prediction of protein–protein interaction network using a multi-objective optimization approach

Protein-Protein Interactions (PPIs) are very important as they coordinate almost all cellular processes. This paper attempts to formulate PPI prediction problem in a multi-objective optimization framework. The scoring functions for the trial solution deal with simultaneous maximization of functional similarity, strength of the domain interaction profiles, and the number of common neighbors of the proteins predicted to be interacting. The above optimization problem is solved using the proposed Firefly Algorithm with Nondominated Sorting. Experiments undertaken reveal that the proposed PPI prediction technique outperforms existing methods, including gene ontology-based Relative Specific Similarity, multi-domain-based Domain Cohesion Coupling method, domain-based Random Decision Forest method, Bagging with REP Tree, and evolutionary/swarm algorithm-based approaches, with respect to sensitivity, specificity, and F1 score.

An Evolutionary Approach for Analysing the Effect of Interaction Site Structural Features on Protein- Protein Complex Formation

Protein-protein complexes that dissociate and associate readily,often depending on the physiological condition or environment,play an important role in many biological processes.The impact of the features responsible for protein complex formation is not uniform. In this paper we have tried to rank the features required for stable protein-protein complex formation.We have employed Artificial Bee Colony with Temporal Difference Q learning algorithm to assign weights to the various atomic structure features. Experiments with data provide evidence that such an approach leads to improved clustering performance.

Evolutionary approach for designing protein-protein interaction network using artificial bee colony optimization

In the new paradigm for studying biological phenomena represented by System Biology, cellular components are not considered in isolation but as forming complex networks of relationships. Protein-Protein Interaction (PPI) networks are among the first objects studied from this new point of view. The paper addresses an interesting approach to protein-protein interaction problem using Artificial Bee Colony (ABC) optimization algorithm. In this work, PPI is formulated as an optimization problem. The binding energy and mismatch in phylogenetic profiles of two bound proteins are used as a scoring function for the solutions. Results are demonstrated for three different networks both numerically and pictorially. Experimental results reveal that the proposed method outperforms Differential Evolution (DE) based PPI network design method considering the intra- and inter-molecular energies of the evolved molecules and the phylogenetic profiles of the proteins in the network.

A meta-heuristic approach to predict protein-protein interaction network

This paper formulates the protein-protein interaction (PPI) prediction problem as a multi-objective optimization (MOO) problem. The focus here is to jointly maximize i) the number of common neighbors of the proteins predicted to be interacting, ii) their functional similarity, and iii) the ratio between their individual accessible solvent area and that of the corresponding protein-protein complex. The above MOO problem is solved using a fusion of the differential evolution for multi-objective optimization and the stochastic learning automata. Experiments undertaken reveal that the proposed PPI prediction technique outperforms existing methods with respect to sensitivity, specificity, and F1 score.