Structural characterization of the interaction between histone tails and effector modules (bromodomains, chromodomains, PHD fingers, etc.) is fundamental to understand the mechanistic aspects of epigenetic regulation of gene expression. In recent years many researchers have applied this approach to specific systems, thus providing a valuable but fragmentary view of the histone-effector interaction. In our work we use this information to characterize the structural features of the two main components of this interaction, histone peptides and the binding site of effector domains (focusing on those which target modified lysines), and increase our knowledge on its specificity determinants. Our results show that the binding sites of effectors are structurally variable, but some clear trends allow their classification in three main groups: flat-groove, narrow-groove and cavity-insertion. In addition, we found that even within these classes binding site variability is substantial. These results in context with the work from other researchers indicate that the there are at least two determinants of binding specificity in the binding site of effector modules. Finally, our analysis of the histone peptides sheds light on the structural transition experienced by histone tails upon effector binding, showing that it may vary depending on the local properties of the sequence stretch considered, thus allowing us to identify an additional specificity determinant for this interaction. Overall, the results of our analysis contribute to clarify the origins of specificity: different regions of the binding site and, in particular, differences in the disorder-order transitions experienced by different histone sequence stretches upon binding.

Big Data and the IoT explosion has made clustering multivariate Time Series (TS) one of the most effervescent research fields. From Bio-informatics to Business and Management, multivariate TS are becoming more and more interesting as they allow to match events the co-occur in time but that is hardly noticeable. This study represents a step forward in our research. We firstly made use of Recurrent Neural Networks and transfer learning to analyze each example, measuring similarities between variables. All the results are finally aggregated to create an adjacency matrix that allows extracting the groups. In this second approach, splines are introduced to smooth the TS before modeling; also, this step avoid to learn from data with high variation or with noise. In the experiments, the two solutions are compared suing the same proof-of-concept experimentation.

No abstract

No abstract

No abstract