Current Bioinformatics - Volume 6, Issue 4, 2011

Protein-protein interactions (PPIs) have been marked as the main actors for all of the processes taking place in a cell and therefore great efforts have been made towards the understanding of their biological function. Today, new highthroughput technologies generate vast amounts of interaction data even with a limited number of experiments. The analysis of these data can lead to valuable conclusions about the cell organization such as protein complex detection, characterization of protein function, identification of protein pathways etc. Various techniques have been applied to analyze PPI data based on different strategies. Web interfaces that have been developed to host these methods consist of valuable tools which are often available to all users. In this review, we describe a collection of such web tools to analyze PPI data and more that are applicable to a wider range of problems. The functionality of each web tool is described as well as their compatibility with other resources. An overview of the technologies that are supported by such tools for their activities is also provided.

Protein-Protein Interactions (PPIs) play a very important role in many cellular processes and a variety of experimental approaches have been developed for their identification. These approaches however are partial, timeconsuming and they usually suffer from high error rates. Recently, computational methods have been employed to assist for the prediction producing encouraging results. With this work we offer a critical review of recent computational PPI prediction methods by evaluating their strengths and limitations. Moreover we discuss open problems common to all schemes and try to suggest solutions. Finally we propose future research directions which could potentially more effectively handle some of the restrictions of existing approaches.

Influenza A surface glycoproteins, hemagglutinin (HA) and neuraminidase (NA), are the major targets of neutralizing antibodies which necessitate their inclusion in influenza vaccines. The frequent antigenic drift and shift of these proteins are responsible for the annual epidemics and occasional pandemics of influenza. Therefore, vaccines must be reformulated annually to include the HA and NA proteins of the viral strains predicted for the upcoming flu season. There are inherent limitations in the annual vaccine preparation process since testing for effectiveness, approval by regulatory agencies and distribution of these vaccines can take approximately 6 months. These drawbacks are most pronounced in situations of pandemic influenza such as the recent swine-origin 2009 H1N1 pandemic, thus there is a critical need for the development of new antiviral and vaccine strategies. One promising new area of research is targeting the conserved regions of influenza surface glycoproteins. Recent studies have demonstrated that vaccines based on highly conserved sequences or antibodies raised against conserved epitopes can protect animals against diverse influenza A virus subtypes. In this review, we focus on the challenges associated with the two main surface glycoproteins, HA and NA, while emphasizing recent advances in bioinformatics tools and their contribution to the design of new diagnostics, vaccines and antivirals.

Microarray technology is capable of providing biomedical and biological researchers with a massive amount of gene expression information to enable rapid significant discoveries in life sciences. Microarray data analysis has been developing at a fast pace during the last decade and has become a popular and standard research method for gene expression studies undertaken by genomics research groups worldwide. Many computational tools have been applied to mine this data in order to discover biologically meaningful knowledge. One of the most useful analysis tools is the fuzzy clustering approach which can be modeled in many types of the continuous partitions of data and are well known for its ability to identify co-expressed genes and annotate functions for novel genes. As the computational analysis of microarray data has been developing rapidly, articles surveying its progress of research and developments are periodically needed. In this paper, we review the recent research into microarray data analysis based on fuzzy clustering algorithms and present a newly developed fuzzy clustering technique which, potentially, can be applied to perform microarray data analysis.

The rapid growth of the Protein Data Bank (PDB) highlights a great challenge for researchers to predict the binding sites of protein for specific metal ion(s). Experimental determination of functional features of a protein is expensive, time consuming and difficult to automate. Therefore, there is a great demand of computational methods for predicting functional features of protein. This review sheds light on currently available in-silico methods including different tools and databases which are based on various information of metal ion and their binding sites (protein residue length, amino acid composition, geometrical and molecular information etc.) and determines the efficiency, speed and accuracy by using diverse algorithms which make the tools beneficial.

Proteins play a major role in biochemical and biophysical features of living organisms and the knowledge of their function is crucial in the development of new drugs, agriculture productivity improvement, biofuel production and several industrial products derived biologically. In current annotation approach, functional prediction of unknown proteins such as hypothetical proteins present in various genomes is a challenging task to the biological community. This incomplete genome information causes gaps in the knowledge especially in the area of drug discovery and metabolic engineering. Hence, an integrated genome-scale re-annotation has been evolved, which is the most promising approach to predict the functions of unknown or hypothetical proteins. In this approach, along with BLAST many other tools output will be integrated for better understanding the function of a given protein sequence. In this review, we describe the integrated re-annotation approach methods that will be helpful in systems level study of microorganisms.

The discovery of microRNAs (miRNAs), about a decade ago, has completely changed our understanding of the complexity of gene regulatory networks. It has already been shown that they are abundantly found in many organisms and can regulate hundreds of genes in post-transcriptional level. To elucidate the individual or co-operative effects of miRNAs, it is required to place them in the overall network of gene regulation and link them to other pathways and systems-level processes. One key step in this effort is predicting targets of individual miRNAs. Although current tools are helpful in predicting miRNA-mRNA binding to a considerable extent, they are not able to model many-to-many relationships between miRNAs and their targets using solely sequence information. Therefore, other types of information sources have been employed for better prediction of these functional relationships. This report focuses on the state-of-theart solutions and current challenges on mining miRNA-related data to discover the systems-level role of miRNAs, with an emphasis on the integration of different information sources. We aim to provide new insights for fusion of different types of biochemical and experimental information sources which may facilitate functional analysis of miRNAs.