Simultaneous inference in general parametric models

Simultaneous inference is a common problem in many areas of application. If multiple null hypotheses are tested simultaneously, the probability of rejecting erroneously at least one of them increases beyond the pre-specified significance level. Simultaneous inference procedures have to be used which adjust for multiplicity and thus control the overall type I error rate. In this paper we describe simultaneous inference procedures in general parametric models, where the experimental questions are specified through a linear combination of elemental model parameters. The framework described here is quite general and extends the canonical theory of multiple comparison procedures in ANOVA models to linear regression problems, generalized linear models, linear mixed effects models, the Cox model, robust linear models, etc. Several examples using a variety of different statistical models illustrate the breadth of the results. For the analyses we use the R add-on package multcomp, which provides a convenient interface to the general approach adopted here.     

Setup of a scientific computing environment for computational biology: Simulation of a genome-scale metabolic model of Escherichia coli as an example

Computational analysis of biological data is becoming increasingly important, especially in this era of big data. Computational analysis of biological data allows efficiently deriving biological insights for given data, and sometimes even counterintuitive ones that may challenge the existing knowledge. Among experimental researchers without any prior exposure to computer programming, computational analysis of biological data has often been considered to be a task reserved for computational biologists. However, thanks to the increasing availability of user-friendly computational resources, experimental researchers can now easily access computational resources, including a scientific computing environment and packages necessary for data analysis. In this regard, we here describe the process of accessing Jupyter Notebook, the most popular Python coding environment, to conduct computational biology. Python is currently a mainstream programming language for biology and biotechnology. In particular, Anaconda and Google Colaboratory are introduced as two representative options to easily launch Jupyter Notebook. Finally, a Python package COBRApy is demonstrated as an example to simulate 1) specific growth rate of Escherichia coli as well as compounds consumed or generated under a minimal medium with glucose as a sole carbon source, and 2) theoretical production yield of succinic acid, an industrially important chemical, using E. coli. This protocol should serve as a guide for further extended computational analyses of biological data for experimental researchers without computational background.     

Structural models for dimerization of G-protein coupled receptors: the opioid receptor homodimers

Among the most exciting functional features of G-protein coupled receptors (GPCRs) that are coming into focus lately are those relating to the role and structural characteristics of their oligomerization (mostly homo- and heterodimers). The structural underpinnings of these novel functional insights are still not clear, as current experimental techniques have not yet succeeded in identifying the dimerization interfaces between GPCR monomers. Two computational approaches have recently been designed in our lab to provide reasonable three-dimensional (3D) molecular models of the transmembrane (TM) regions of GPCR dimers based on a combination of the structural information of receptor monomers and analyses of correlated mutations in receptor families. The modeling of GPCR heterodimers has been described recently. We present here a related approach for modeling of GPCR homodimers that identifies the interfaces in the most likely configurations of the complexes. The approach is illustrated for the three cloned opioid receptor subtypes (OPRD, OPRM, and OPRK).     

Structure-function analysis of the coxsackievirus protein 3A: identification of residues important for dimerization, viral rna replication, and transport inhibition

The coxsackievirus B3 3A protein forms homodimers and plays important roles in both viral RNA (vRNA) replication and the viral inhibition of intracellular protein transport. The molecular determinants that are required for each of these functions are yet poorly understood. Based on the NMR structure of the closely related poliovirus 3A protein, a molecular model of the coxsackievirus B3 3A protein was constructed. Using this structural model, specific mutants were designed to study the structure-function relationship of 3A. The mutants were tested for their capacity to dimerize, support vRNA replication, and block protein transport. A hydrophobic interaction between the monomers and an intermolecular salt bridge were identified as major determinants required for dimerization. We show that dimerization is important for both efficient vRNA replication and inhibition of protein transport. In addition, determinants were identified that were not required for dimerization but that were essential for either one of the biological functions of 3A. The combination of the in silico and in vivo results obtained in this study provides important insights in both the structural and functional aspects of 3A.     

Inferring network interactions within a cell

The continuing growth in high-throughput data acquisition has led to a proliferation of network models to represent and analyse biological systems. These networks involve distinct interaction types detected by a combination of methods, ranging from directly observed physical interactions based in biochemistry to interactions inferred from phenotype measurements, genomic expression and comparative genomics. The discovery of interactions increasingly requires a blend of experimental and computational methods. Considering yeast as a model system, recent analytical methods are reviewed here and specific aims are proposed to improve network interaction inference and facilitate predictive biological modelling.     

Probing fibroblast growth factor dimerization and role of heparin-like glycosaminoglycans in modulating dimerization and signaling

For a number of growth factors and cytokines, ligand dimerization is believed to be central to the formation of an active signaling complex. In the case of fibroblast growth factor-2 (FGF2) signaling, heparin/heparan sulfate-like glycosaminoglycans (HLGAGs) are involved through interaction with both FGF2 and its receptors (FGFRs) in assembling a tertiary complex and modulating FGF2 activity. Biochemical data have suggested different modes of HLGAG-induced FGF2 dimerization involving specific protein-protein contacts. In addition, several recent x-ray crystallography studies of FGF.FGFR and FGF.FGFR.HLGAG complexes have revealed other modes of molecular assemblage, with no FGF-FGF contacts. All these different biochemical and structural findings have clarified less and in fact raised more questions as to which mode of FGF2 dimerization, if any, is essential for signaling. In this study, we address the issue of FGF2 dimerization in signaling using a combination of biochemical, biophysical, and site-directed mutagenesis approaches. Our findings presented here provide direct evidence of FGF2 dimerization in mediating FGF2 signaling.     

Towards the prediction of essential genes by integration of network topology, cellular localization and biological process information

Background  

The identification of essential genes is important for the understanding of the minimal requirements for cellular life and for practical purposes, such as drug design. However, the experimental techniques for essential genes discovery are labor-intensive and time-consuming. Considering these experimental constraints, a computational approach capable of accurately predicting essential genes would be of great value. We therefore present here a machine learning-based computational approach relying on network topological features, cellular localization and biological process information for prediction of essential genes.     

ComPath: comparative enzyme analysis and annotation in pathway/subsystem contexts

Background  

Once a new genome is sequenced, one of the important questions is to determine the presence and absence of biological pathways. Analysis of biological pathways in a genome is a complicated task since a number of biological entities are involved in pathways and biological pathways in different organisms are not identical. Computational pathway identification and analysis thus involves a number of computational tools and databases and typically done in comparison with pathways in other organisms. This computational requirement is much beyond the capability of biologists, so information systems for reconstructing, annotating, and analyzing biological pathways are much needed. We introduce a new comparative pathway analysis workbench, ComPath, which integrates various resources and computational tools using an interactive spreadsheet-style web interface for reliable pathway analyses.     

Identifying Cytochrome P450 Functional Networks and Their Allosteric Regulatory Elements

Cytochrome P450 (CYP) enzymes play key roles in drug metabolism and adverse drug-drug interactions. Despite tremendous efforts in the past decades, essential questions regarding the function and activity of CYPs remain unanswered. Here, we used a combination of sequence-based co-evolutionary analysis and structure-based anisotropic thermal diffusion (ATD) molecular dynamics simulations to detect allosteric networks of amino acid residues and characterize their biological and molecular functions. We investigated four CYP subfamilies (CYP1A, CYP2D, CYP2C, and CYP3A) that are involved in 90% of all metabolic drug transformations and identified four amino acid interaction networks associated with specific CYP functionalities, i.e., membrane binding, heme binding, catalytic activity, and dimerization. Interestingly, we did not detect any co-evolved substrate-binding network, suggesting that substrate recognition is specific for each subfamily. Analysis of the membrane binding networks revealed that different CYP proteins adopt different membrane-bound orientations, consistent with the differing substrate preference for each isoform. The catalytic networks were associated with conservation of catalytic function among CYP isoforms, whereas the dimerization network was specific to different CYP isoforms. We further applied low-temperature ATD simulations to verify proposed allosteric sites associated with the heme-binding network and their role in regulating metabolic fate. Our approach allowed for a broad characterization of CYP properties, such as membrane interactions, catalytic mechanisms, dimerization, and linking these to groups of residues that can serve as allosteric regulators. The presented combined co-evolutionary analysis and ATD simulation approach is also generally applicable to other biological systems where allostery plays a role.     

Emerging technologies: systems biology

The field of systems biology is based on the paradigm that the whole is greater than the sum of the parts. Through a combination of high-throughput experiments analyzing "-omic" scale phenomenon and the development of new computational techniques and algorithms, it is now feasible to study biological systems in a way that was previously not possible. During the 232nd National Meeting of the American Chemical Society, a session devoted to the emerging technology of Systems Biology was held. A number of talks on a wide variety of subjects covering cell signaling, network regulation and analysis, novel experimental procedures, synthetic biology, and metabolic flux analysis were presented. All of these approaches shared the common theme of using a systems biology approach to aid in the understanding of fundamental biology, with an eye toward applications for the benefit of society.     

Single-case designs and clinical biofeedback experimentation

Rigorous evaluation of the effects of biofeedback with clinical populations is necessary, but practical problems often preclude utilization of between-groups experimental designs involving large numbers of clients with clinically relevant problems. Single-case experimental designs provide a viable alternative for answering most research questions. In addition, single-case designs possess several distinct advantages for biofeedback research, including a focus on clinical significance, the use of variability as data not error, unique procedures for establishing generality of findings, and an ability to deal with ethical concerns in clinical research. Basic procedures in the use of single-case experimental designs are described and illustrations in clinical biofeedback research are provided.     

Perspectives on the automatic design of regulatory systems for synthetic biology

Automatic design is based on computational modeling and optimization methods to provide prototype designs to targeted problems in an unsupervised manner. For biological circuits, we need to produce quantitative predictions of cell behavior for a given genotype as consequence of the different molecular interactions. Automatic design techniques aim at solving the inverse problem of finding the sequences of nucleotides that better fit a targeted behavior. In the post-genomic era, our molecular knowledge and modeling capabilities have allowed to start using such methodologies with success. Herein, we describe how the emergence of this new type of tools could enable novel synthetic biology applications. We highlight the essential elements to develop automatic design procedures for synthetic biology pointing out their advantages and bottlenecks. We discuss in detail the experimental difficulties to overcome in the in vivo implementation of designed networks. The use of automatic design to engineer biological networks is starting to emerge as a new technique to perform synthetic biology, which should not be neglected in the future.     

Dose response problems in carcinogenesis.

The estimation of risks from exposure to carcinogens is an important problem from the viewpoint of protection of human health. It also poses some very difficult dose-response problems. Two dose-response models may fit experimental data about equally well and yet predict responses that differ by many orders of magnitude at low doses. Mechanisms of carcinogenesis are not sufficiently understood so that the shape of the dose-response curve at low doses can be satisfactorily predicted. Mathematical theories of carcinogenesis and statistical procedures can be of use with dose-reponse problems such as this and, in addition, can lead to a better understanding of the mechanisms of carcinogenesis. In this paper, mathematical dose-response models of carcinogenesis are considered as well as various proposed dose-response procedures for estimating carcinogenic risks at low doses. Areas are suggested in which further work may be useful. These areas include experimental design problems, statistical procedures for use with time-to-occurrence data, and mathematical models that incorporate such biological features as pharmacokinetics of carcinogens, synergistic effects, DNA repair, susceptible subpopulations, and immune reactions.     

GUIMACS - a Java based front end for GROMACS

Molecular dynamics simulations have gained importance due to their ability to provide valuable insights into understanding structure-function relationships of biological macromolecules. With increasing computational speeds there has been a substantial demand for optimization of simulation algorithms to obtain results even faster. With this on one hand, the need for ease of operation lies on the other. GUI front end programs are important appurtenances to ease the use of command line programs. Effective use of command line based programs requires basic knowledge of the UNIX shell and at least one of the UNIX based text editors, making it difficult for pure biologists to use them efficiently. GROMACS, a widely used suite of molecular dynamics simulation and analysis programs, is no exception to this. As a matter of fact, the increasing dependency of experimental procedures on computational methods for accentuating certain key experimental findings increases the need for interactivity in use of command-line based packages.     

Analysis of the quaternary structure of the MutL C-terminal domain

The dimeric DNA mismatch repair protein MutL has a key function in communicating mismatch recognition by MutS to downstream repair processes. Dimerization of MutL is mediated by the C-terminal domain, while activity of the protein is modulated by the ATP-dependent dimerization of the highly conserved N-terminal domain. Recently, a crystal structure analysis of the Escherichia coli MutL C-terminal dimerization domain has been reported and a model for the biological dimer was proposed. In this model, dimerization is mediated by the internal (In) subdomain comprising residues 475-569. Here, we report a computational analysis of all protein interfaces observed in the crystal structure and suggest that the biological dimer interface is formed by a hydrophobic surface patch of the external (Ex) subdomain (residues 432-474 and 570-615). Moreover, sequence analysis revealed that this surface patch is conserved among the MutL proteins. To test this hypothesis, single and double-cysteine variants of MutL were generated and tested for their ability to be cross-linked with chemical cross-linkers of various size. Finally, deletion of the C-terminal residues 605-615 abolished homodimerization. The biochemical data are fully compatible with a revised model for the biological dimer, which has important implications for understanding the heterodimerization of eukaryotic MutL homologues, modeling the MutL holoenzyme and predicting protein-protein interaction sites.     

Analyzing time series gene expression data

MOTIVATION: Time series expression experiments are an increasingly popular method for studying a wide range of biological systems. However, when analyzing these experiments researchers face many new computational challenges. Algorithms that are specifically designed for time series experiments are required so that we can take advantage of their unique features (such as the ability to infer causality from the temporal response pattern) and address the unique problems they raise (e.g. handling the different non-uniform sampling rates). RESULTS: We present a comprehensive review of the current research in time series expression data analysis. We divide the computational challenges into four analysis levels: experimental design, data analysis, pattern recognition and networks. For each of these levels, we discuss computational and biological problems at that level and point out some of the methods that have been proposed to deal with these issues. Many open problems in all these levels are discussed. This review is intended to serve as both, a point of reference for experimental biologists looking for practical solutions for analyzing their data, and a starting point for computer scientists interested in working on the computational problems related to time series expression analysis.     

A Computationally Guided Protein-Interaction Screen Uncovers Coiled-Coil Interactions Involved in Vesicular Trafficking

Mapping protein-protein interactions at a domain or motif level can provide structural annotation of the interactome. The α-helical coiled coil is among the most common protein-interaction motifs, and proteins predicted to contain coiled coils participate in diverse biological processes. Here, we introduce a combined computational/experimental screening strategy that we used to uncover coiled-coil interactions among proteins involved in vesicular trafficking in Saccharomyces cerevisiae. A number of coiled-coil complexes have already been identified and reported to play important roles in this important biological process. We identify additional examples of coiled coils that can form physical associations. The computational strategy used to prioritize coiled-coil candidates for testing dramatically improved the efficiency of discovery in a large experimental screen. As assessed by comprehensive yeast two-hybrid assays, computational prefiltering retained 90% of positive interacting pairs and eliminated > 60% of negatives from a set of interaction candidates. The coiled-coil-mediated interaction network elucidated using the combined computational/experimental approach comprises 80 coiled-coil associations between 58 protein pairs, among which 21 protein interactions have not been previously reported in interaction databases and 26 interactions were previously known at the protein level but have now been localized to the coiled-coil motif. The coiled-coil-mediated interactions were specific rather than promiscuous, and many interactions could be recapitulated in a green fluorescent protein complementation assay. Our method provides an efficient route to discovering new coiled-coil interactions and uncovers a number of associations that may have functional significance for vesicular trafficking.     

Determination of the transition state ensemble for the folding of ubiquitin from a combination of Phi and Psi analyses

Protein engineering techniques have emerged as powerful tools for characterizing transition states (TSs) for protein folding. Recently, the Ψ analysis, in which double-histidine mutations create the possibility of reversible crosslinking in the native state, has been proposed as an additional approach to the well-established Φ analysis. We present here a combination of these two procedures for defining the structure of the TS of ubiquitin, a small α/β protein that has been used extensively as a model system for both experimental and computational studies of the protein-folding process. We performed a series of molecular dynamics simulations in which Φ and Ψ values were used as ensemble-averaged structural restraints to determine an ensemble of structures representing the TS of ubiquitin. Although the available Ψ values for ubiquitin did not, by themselves, generate well-defined TS ensembles, the inclusion of the restricted set of zero or unity values, but not fractional ones, provided useful complementary information to the Φ analysis. Our results show that the TS of ubiquitin is formed by a relatively narrow ensemble of structures exhibiting an overall native-like topology in which the N-terminal and C-terminal regions are in close proximity.