BMC Bioinformatics

Protein phosphorylation networks play an important role in cell signaling. In these networks, phosphorylation of a protein kinase usually leads to its activation, which in turn will phosphorylate its downstream target proteins. A phosphorylation network is essentially a causal network, which can be learned by causal inference algorithms. Prior efforts have applied such algorithms to data measuring protein phosphorylation levels, assuming that the phosphorylation levels represent protein activity states. However, the phosphorylation status of a kinase does not always reflect its activity state, because interventions such as inhibitors or mutations can directly affect its activity state without changing its phosphorylation status. Thus, when cellular systems are subjected to extensive perturbations, the statistical relationships between phosphorylation states of proteins may be disrupted, making it difficult to reconstruct the true protein phosphorylation network. Here, we describe a novel framework to address this challenge. We have developed a causal discovery framework that explicitly represents the activity state of each protein kinase as an unmeasured variable and developed a novel algorithm called “InferA” to infer the protein activity states, which allows us to incorporate the protein phosphorylation level, pharmacological interventions and prior knowledge. We applied our framework to simulated datasets and to a real-world dataset. The simulation experiments demonstrated that explicit representation of activity states of protein kinases allows one to effectively represent the impact of interventions and thus enabled our framework to accurately recover the ground-truth causal network. Results from the real-world dataset showed that the explicit representation of protein activity states allowed an effective and data-driven integration of the prior knowledge by InferA, which further leads to the recovery of a phosphorylation network that is more consistent with experiment results. Explicit representation of the protein activity states by our novel framework significantly enhances causal discovery of protein phosphorylation networks.

Because biological networks exhibit a high-degree of robustness, a systemic understanding of their architecture and function requires an appraisal of the network design principles that confer robustness. In this project, we conduct a computational study of the contribution of three degree-based topological properties (transcription factor-target ratio, degree distribution, cross-talk suppression) and their combinations on the robustness of transcriptional regulatory networks. We seek to quantify the relative degree of robustness conferred by each property (and combination) and also to determine the extent to which these properties alone can explain the robustness observed in transcriptional networks. To study individual properties and their combinations, we generated synthetic, random networks that retained one or more of the three properties with values derived from either the yeast or E. coli gene regulatory networks. Robustness of these networks were estimated through simulation. Our results indicate that the combination of the three properties we considered explains the majority of the structural robustness observed in the real transcriptional networks. Surprisingly, scale-free degree distribution is, overall, a minor contributor to robustness. Instead, most robustness is gained through topological features that limit the complexity of the overall network and increase the transcription factor subnetwork sparsity. Our work demonstrates that (i) different types of robustness are implemented by different topological aspects of the network and (ii) size and sparsity of the transcription factor subnetwork play an important role for robustness induction. Our results are conserved across yeast and E Coli, which suggests that the design principles examined are present within an array of living systems.

Carbohydrates are a class of large and diverse biomolecules, ranging from a simple monosaccharide to large multi-branching glycan structures. The covalent linkage of a carbohydrate to the nitrogen atom of an asparagine, a process referred to as N-linked glycosylation, plays an important role in the physiology of many living organisms. Most software for glycan modeling on a personal desktop computer requires knowledge of molecular dynamics to interface with specialized programs such as CHARMM or AMBER. There are a number of popular web-based tools that are available for modeling glycans (e.g., GLYCAM-WEB (http:// https://dev.glycam.org/gp/ ) or Glycosciences.db ( http://www.glycosciences.de/ )). However, these web-based tools are generally limited to a few canonical glycan conformations and do not allow the user to incorporate glycan modeling into their protein structure modeling workflow. Here, we present Glycosylator, a Python framework for the identification, modeling and modification of glycans in protein structure that can be used directly in a Python script through its application programming interface (API) or through its graphical user interface (GUI). The GUI provides a straightforward two-dimensional (2D) rendering of a glycoprotein that allows for a quick visual inspection of the glycosylation state of all the sequons on a protein structure. Modeled glycans can be further refined by a genetic algorithm for removing clashes and sampling alternative conformations. Glycosylator can also identify specific three-dimensional (3D) glycans on a protein structure using a library of predefined templates. Glycosylator was used to generate models of glycosylated protein without steric clashes. Since the molecular topology is based on the CHARMM force field, new complex sugar moieties can be generated without modifying the internals of the code. Glycosylator provides more functionality for analyzing and modeling glycans than any other available software or webserver at present. Glycosylator will be a valuable tool for the glycoinformatics and biomolecular modeling communities.

Reproducibility in Data Analysis research has long been a significant concern, particularly in the areas of Bioinformatics and Computational Biology. Towards the aim of developing reproducible and reusable processes, Data Analysis management tools can help giving structure and coherence to complex data flows. Nonetheless, improved software quality comes at the cost of additional design and planning effort, which may become impractical in rapidly changing development environments. I propose that an adjustment of focus from processes to data in the management of Bioinformatic pipelines may help improving reproducibility with minimal impact on preexisting development practices. In this paper I introduce the repo R package for bioinformatic analysis management. The tool supports a data-centered philosophy that aims at improving analysis reproducibility and reusability with minimal design overhead. The core of repo lies in its support for easy data storage, retrieval, distribution and annotation. In repo the data analysis flow is derived a posteriori from dependency annotations. The repo package constitutes an unobtrusive data and flow management extension of the R statistical language. Its adoption, together with good development practices, can help improving data analysis management, sharing and reproducibility, especially in the fields of Bioinformatics and Computational Biology.

Gene ontology (GO) enrichment is commonly used for inferring biological meaning from systems biology experiments. However, determining differential GO and pathway enrichment between DNA-binding experiments or using the GO structure to classify experiments has received little attention. Herein, we present a bioinformatics tool, CompGO, for identifying Differentially Enriched Gene Ontologies, called DiEGOs, and pathways, through the use of a z-score derivation of log odds ratios, and visualizing these differences at GO and pathway level. Through public experimental data focused on the cardiac transcription factor NKX2-5, we illustrate the problems associated with comparing GO enrichments between experiments using a simple overlap approach. We have developed an R/Bioconductor package, CompGO, which implements a new statistic normally used in epidemiological studies for performing comparative GO analyses and visualizing comparisons from .BED data containing genomic coordinates as well as gene lists as inputs. We justify the statistic through inclusion of experimental data and compare to the commonly used overlap method. CompGO is freely available as a R/Bioconductor package enabling easy integration into existing pipelines and is available at: http://www.bioconductor.org/packages/release/bioc/html/CompGO.html packages/release/bioc/html/CompGO.html

There has been a trend in increasing the phylogenetic scope of genome sequencing without finishing the sequence of the genome. Increasing numbers of genomes are being published in scaffold or contig form. Rearrangement algorithms, however, including gene order-based phylogenetic tools, require whole genome data on gene order or syntenic block order. How then can we use rearrangement algorithms to compare genomes available in scaffold form only? Can the comparative evidence predict the location of unsequenced genes? Our method involves optimally filling in genes missing from the scaffolds, while incorporating the augmented scaffolds directly into the rearrangement algorithms as if they were chromosomes. This is accomplished by an exact, polynomial-time algorithm. We then correct for the number of extra fusion/fission operations required to make scaffolds comparable to full assemblies. We model the relationship between the ratio of missing genes actually absent from the genome versus merely unsequenced ones, on one hand, and the increase of genomic distance after scaffold filling, on the other. We estimate the parameters of this model through simulations and by comparing the angiosperm genomes Ricinus communis and Vitis vinifera. The algorithm solves the comparison of genomes with 18,300 genes, including 4500 missing from one genome, in less than a minute on a MacBook, putting virtually all genomes within range of the method.

Recognizing the different functional parts of genes, such as promoters, translation initiation sites, donors, acceptors and stop codons, is a fundamental task of many current studies in Bioinformatics. Currently, the most successful methods use powerful classifiers, such as support vector machines with various string kernels. However, with the rapid evolution of our ability to collect genomic information, it has been shown that combining many sources of evidence is fundamental to the success of any recognition task. With the advent of next-generation sequencing, the number of available genomes is increasing very rapidly. Thus, methods for making use of such large amounts of information are needed. In this paper, we present a methodology for combining tens or even hundreds of different classifiers for an improved performance. Our approach can include almost a limitless number of sources of evidence. We can use the evidence for the prediction of sites in a certain species, such as human, or other species as needed. This approach can be used for any of the functional recognition tasks cited above. However, to provide the necessary focus, we have tested our approach in two functional recognition tasks: translation initiation site and stop codon recognition. We have used the entire human genome as a target and another 20 species as sources of evidence and tested our method on five different human chromosomes. The proposed method achieves better accuracy than the best state-of-the-art method both in terms of the geometric mean of the specificity and sensitivity and the area under the receiver operating characteristic and precision recall curves. Furthermore, our approach shows a more principled way for selecting the best genomes to be combined for a given recognition task. Our approach has proven to be a powerful tool for improving the performance of functional site recognition, and it is a useful method for combining many sources of evidence for any recognition task in Bioinformatics. The results also show that the common approach of heuristically choosing the species to be used as source of evidence can be improved because the best combinations of genomes for recognition were those not usually selected. Although the experiments were performed for translation initiation site and stop codon recognition, any other recognition task may benefit from our methodology.