Comparison of parallel multi-objective approaches to protein structure prediction

Protein structure prediction (PSP) is an open problem with many useful applications in disciplines such as medicine, biology and biochemistry. As this problem presents a vast search space and the analysis of each protein structure requires a significant amount of computing time, it is necessary to take advantage of high-performance parallel computing platforms as well as to define efficient search procedures in the space of possible protein conformations. In this paper we compare two parallel procedures for PSP which are based on different multi-objective optimization approaches, i.e. PAES (Knowles and Corne in Proc. Congr. Evol. Comput. 1:98–105, 1999) and NSGA2 (Deb et al. in IEEE Trans. Evol. Comput. 6:182–197, 2002). Although both procedures include techniques to take advantage of known protein structures and strategies to simplify the search space through the so-called rotamer library and adaptive mutation operators, they present different profiles with respect to their implicit parallelism.     

Dynamic protein–protein interaction networks construction using firefly algorithm

Protein–protein interaction (PPI) networks are dynamic in the real world. That is, at different times and under different conditions, the interaction among proteins may or may not be active. In different dataset, PPI networks might be gathered as static or dynamic networks. For the conversion of static PPI networks to time graphs, i.e., dynamic PPI networks, additional information like gene expression and gene co-expression profiles is used. One of the challenges in system biology is to determine appropriate thresholds for converting static PPI networks to dynamic PPI networks based on active proteins. In the available methods, fixed thresholds are used for all genes. However, the purpose of this study is to determine an adaptive unique threshold for each gene. In this study, the available additional information at different times and conditions and gold-standard protein complexes was employed to determine fitting thresholds. By so doing, the problem is converted into an optimization problem. Thereafter, the problem is solved using the firefly meta-heuristic optimization algorithm. One of the most remarkable aspects of this study is determining the attractiveness function in the firefly algorithm. In this study, attraction is defined as a combination of standard complexes and gene co-expressions. Then, active proteins are specified utilizing the created thresholds. The MCL, ClusterOne, MCODE and Coach algorithms are used for final evaluation. The experimental results about BioGRID dataset and CYC2008 gold-standard protein complexes indicated that the produced dynamic PPI networks by the proposed method have better results than the earlier methods.     

GPU acceleration of Fitch’s parsimony on protein data: from Kepler to Turing

The analysis of complex biological datasets beyond DNA scenarios is gaining increasing interest in current bioinformatics. Particularly, protein sequence data introduce additional complexity layers that impose new challenges from a computational perspective. This work is aimed at investigating GPU solutions to address these issues in a representative algorithm from the phylogenetics field: Fitch’s parsimony. GPU strategies are adopted in accordance with the protein-based formulation of the problem, defining an optimized kernel that takes advantage of data parallelism at the calculations associated with different amino acids. In order to understand the relationship between problem sizes and GPU capabilities, an extensive evaluation on a wide range of GPUs is conducted, covering all the recent NVIDIA GPU architectures—from Kepler to Turing. Experimental results on five real-world datasets point out the benefits that imply the exploitation of state-of-the-art GPUs, representing a fitting approach to address the increasing hardness of protein sequence datasets.

     

Congo Red bound to α-1-proteinase inhibitor as a model of supramolecular ligand and protein complex

The complex formation and structure of α-1-proteinase inhibitor with supramolecular ligand Congo Red was predicted using molecular mechanics and molecular dynamics simulation. A seven-molecule Congo Red ligand was introduced to the cleft in β-sheet “A” of an α-1-proteinase inhibitor in place of the peptide chain fragment (342-358) which occupies this locus in the cleaved form of the inhibitor. The striking similarity of Congo Red and peptide chain (342-358) insertion effects, observed by comparison of root mean square (r.m.s.)–distance plots as protein stability increased, confirmed the reliability of the constructed complex. The binding predicted theoretically for the one available cleft in the β-sheet, limited to a few Congo Red molecules, was verified experimentally. α-1-proteinase inhibitor was chosen for this study because of the known natural instability of its β-pleated sheet, but the model is believed to represent other Congo Red complexes involving proteins whose accessibility for dye penetration may be triggered by function-derived structural alterations or may be generated in unfolding conditions.     

Gene expression and protein–protein interaction data for identification of colon cancer related genes using <Emphasis Type="Italic">f</Emphasis>-information measures

One of the most important and challenging problems in functional genomics is how to select the disease genes. In this regard, the paper presents a new computational method to identify disease genes. It judiciously integrates the information of gene expression profiles and shortest path analysis of protein–protein interaction networks. While the \(f\)-information based maximum relevance-maximum significance framework is used to select differentially expressed genes as disease genes using gene expression profiles, the functional protein association network is used to study the mechanism of diseases. An important finding is that some \(f\)-information measures are shown to be effective for selecting relevant and significant genes from microarray data. Extensive experimental study on colorectal cancer establishes the fact that the genes identified by the integrated method have more colorectal cancer genes than the genes identified from the gene expression profiles alone, irrespective of any gene selection algorithm. Also, these genes have greater functional similarity with the reported colorectal cancer genes than the genes identified from the gene expression profiles alone. The enrichment analysis of the obtained genes reveals to be associated with some of the important KEGG pathways. All these results indicate that the integrated method is quite promising and may become a useful tool for identifying disease genes.     

Exploring protein–ligand recognition with Binding MOAD

We have recently announced the largest database of protein–ligand complexes, Binding MOAD (Mother of All Databases). After the August 2004 update, Binding MOAD contains 6816 complexes. There are 2220 protein families and 3316 unique ligands. After searching 6000+ crystallography papers, we have obtained binding data for 1793 (27%) of the complexes. We have also created a non-redundant set of complexes with only one complex from each protein family; in that set, 630 (28%) of the unique complexes have binding data. Here, we present information about the data provided at the Binding MOAD website. We also present the results of mining Binding MOAD to map the degree of solvent exposure for binding sites. We have determined that most cavities and ligands (70–85%) are well buried in the complexes. This fits with the common paradigm that a large degree of contact between the ligand and protein is significant in molecular recognition. GoCAV and the GoCAVviewer are the tools we created for this study. To share our data and make our online dataset more useful to other research groups, we have integrated the viewer into the Binding MOAD website (www.BindingMOAD.org).     

Using thermodynamic integration MD simulation to compute relative protein–ligand binding free energy of a GSK3β kinase inhibitor and its analogs

Thermodynamic integration molecular dynamics simulation was used to investigate how TI-MD simulation preforms in reproducing relative protein–ligand binding free energy of a pair of analogous GSK3β kinase inhibitors of available experimental data (see Fig. 1), and to predict the affinity for other analogs. The computation for the pair gave a ΔΔG of 1.0 kcal/mol, which was in reasonably good agreement with the experimental value of −0.1 kcal/mol. The error bar was estimated at 0.5 kcal/mol. Subsequently, we employed the same protocol to proceed with simulations to find analogous inhibitors with a stronger affinity. Four analogs with a substitution at one site inside the binding pocket were the first to be tried, but no significant enhancement in affinity was found. Subsequent simulations for another 7 analogs was focused on substitutions at the benzene ring of another site, which gave two analogs (analogs 9 and 10) with ΔΔG values of −0.6 and −0.8 kcal/mol, respectively. Both analogs had a OH group at the meta position and another OH group at the ortho position at the other side of the benzene ring, as shown in Table 3. To explore further, another 4 analogs with this characteristic were investigated. Three analogs with ΔΔG values of −2.2, −1.7 and −1.2 kcal/mol, respectively, were found. Hydrogen bond analysis suggested that the additional hydrogen bonds of the added OH groups with Gln185 and/or Asn64, which did not appear in the reference inhibitor or as an analog with one substitution only in the examined cases, were the main contributors to an enhanced affinity. A prediction for better inhibitors should interest experimentalists of enzyme and/or cell assays. Analysis of the interactions between GSK3β kinase and the investigated analogs will be useful in the design of GSK3β kinase inhibitors for compounds of this class.     

Some applications of matrices to location of zeros of polynomials†

In previous papers classical theorems on location of zeros of a polynomial with respect to the left half plane Γ₁ or the unit circle Γ₂ have been reformulated more simply in terms of appropriate companion matrices. It is shown how this work can be extended to the problem of zero location with respect to more general regions Γ of the complex plane. The first approach is to apply the bilinear transformation to the given polynomial, so that for example Γ₁ can be mapped into Γ₂ and a matrix representation of this is derived. An alternative method is discussed which relies on transformation of Γ into Γ₂. Some examples illustrate how any theorem involving Hurwitz-typc minors can be expressed in companion matrix terms, with a consequent halving of the orders of the determinants involved.     

Measurement of the Kernels of a Non-linear System of Finite Order†

A method of measuring the Volterra kernels of a finite-order non-linear system is presented. The kernels are obtained individually as a multi-dimensional impulse response. The multi-dimensional kernel transforms also can be obtained by the method described. As an extension, a technique of obtaining the Volterra kernels from a multi-dimensional step response is presented. This technique is useful for non-linear systems which can be considered to be of a given finite order for only a limited range of input amplitudes.     

The Implementation of Graphic Display of FAS

The principle of the fire alarm system and the implementation techniques of its computer graphic display are presented. With the help of the designing thoughts and technology of OOP, a Fire Alarm System (FAS) has been developed by using visual C++ language. The system has been put into use in many important buildings and has played an important role in monitoring and controlling the fire.     

Visualization of curvature plots and evaluation of fairness: an analysis of the effect of ‘scaling’

Changing the aspect-ratio (scale) of a graphical plot significantly affects the “messages” getting across to a viewer. This is true also for curvature plots and other graphs currently used in Curve & Surface Modeling to evaluate fairness. In particular, the sharp corners appearing in such graphs, employed as local fairness indicators by current practices, are significantly affected by nonuniform scaling. A detailed study of this phenomenon is presented offering specific guidelines for correctly “reading” curvature plots.     

Contracted product of hypermatrices via STP of matrices

An equivalent definition of hypermatrices is introduced. The matrix expression of hypermatrices is proposed. Using permutation matrices, the conversion between different matrix expressions is revealed. The various kinds of contracted products of hypermatrices are realized by semi-tensor products (STP) of matrices via matrix expressions of hypermatrices.     

More Proofs of an Axiom of Łukasiewicz

This paper reports results and some new problems in one of the domains to which automatic first-order theorem provers have been most successfully applied: axiomatics of nonclassical propositional logics. It is well known that one of the standard axioms of the denumerable-valued pure implication logic of ukasiewicz becomes derivable from the remainder in the presence of negation. Here it is shown that the same axiom is similarly derivable using conjunction and disjunction instead of negation. This closes a problem left open by Harris and Fitelson (Journal of Automated Reasoning 27, 2001). Related problems are discussed, and five such open problems are proposed as challenges to the automated reasoning community.     

On the Equivalence of Some Models of Computation

In[1],a definition of computation was given.In[3],we have clarified the necessity of thedefinition.Here we prove the equivalence between the definition and some models ofcomputation.Hence the sufficiency of the definition is clear abundantly.     

Evaluation of Fourier transform of impulse response of linear time invariant systems†

Evaluation of the Fourior transform of the system impulse response is an important aspect of the design of control systems. A method suggested hero requires only one cycle of sine or cosine wave to be applied as an input to the system. It is proved that the sum of the sampled values of the output response, at the sampling interval equal to the period of the input wave, directly yields the sign and cosine transforms respectively. The procedure is generalized to any number of complete cycles of input wave, as well as to n/2 cycles whore n is any odd positive integer.