Predicting nucleosome positioning in genomes: Physical and bioinformatic approaches

Abstract

In eukaryotic genomes, nucleosomes are responsible for packaging DNA and controlling gene expression. For this reason, an increasing interest is arising on computational methods capable of predicting the nucleosome positioning along genomes. In this review we describe and compare bioinformatic and physical approaches adopted to predict nucleosome occupancy along genomes. Computational analyses attempt at decoding the experimental nucleosome maps of genomes in terms of certain dinucleotide step periodicity observed along DNA. Such investigations show that highly significant information about the occurrence of a nucleosome along DNA is intrinsic in certain features of the sequence suggesting that DNA of eukaryotic genomes encodes nucleosome organization. Besides the bioinformatic approaches, physical models were proposed based on the sequence dependent conformational features of the DNA chain, which govern the free energy needed to transform recurrent DNA tracts along the genome into the nucleosomal shape. Graphical abstractHighlights > Theoretical models for predicting nucleosome positioning along genomes are reviewed. > Bioinformatics and physical approaches are compared and discussed. > Mechanical properties of DNA rule thermodynamic stability of nucleosomes. > Nucleosomes are regularly spaced along genomes as required for chromatin fibers. KeywordsThermodynamic stability of nucleosomes; nucleosome code; nucleosome distribution along genomes; bioinformatic approaches for the prediction of nucleosome occupancy; physical models predicting nucleosome occupancy.

Coupling between Sequence-Mediated Nucleosome Organization and Genome Evolution

“…Indeed, the sequence-directed nucleosome positioning is directly observed from in vitro data, because chromatin is reconstituted from DNA and histones only, without any other external factors such as remodelers found in vivo. Accordingly, models established from in vitro genome-wide reconstitution of chromatin predict rather well in vivo nucleosome positioning [22,53,[55][56][57][58][59][60], corroborating the hypothesis that the DNA sequence plays a major role among the different factors influencing the position of nucleosomes [61]. During the past 40 years, attempts to describe the sequence-directed nucleosomal positioning showed that one needs to consider two types of mechanisms (Figure 1): (i) positioning mechanisms where DNA motifs at specific location accommodate DNA wrapping in the nucleosome, for example by favoring certain dinucleotides at contact points between DNA and histones; and (ii) inhibiting mechanisms, with sequences such as poly(dA:dT) preventing nucleosome formation [1].…”

“…Nucleosome occupancy encoded in the sequence can presumably be predicted through sequence-based modeling. This was achieved using mainly two types of approaches: bioinformatic models relying on machine learning [22,53,56,58], and physical models relying on energy calculations [55,57,59,60,77]. The general idea of the bioinformatic models is to detect, genome-wide, the sequence features associated with nucleosomal positioning.…”

“…It was linked to the high level of conservation of histones between organisms, whereas nucleoid-associated proteins are more variable, possibly allowing wider range for genomic GC content between species [78]. The physical modeling approach was considered independently by different groups [55,57,60,79]. It is based on intrinsic bending properties of the DNA and thus, its ability to be wrapped around histone octamers.…”

A deformation energy-based model for predicting nucleosome dyads and occupancy

“…In recent years, experimental mapping of genome-wide nucleosome organization has been obtained for several model systems 33 34 35 36 37 38 , such as Saccharomyces cerevisiae , Caenorhabditis elegans , Drosophila melanogaster and Homo sapiens , but the mechanism of nucleosome positioning still remains elusive. A variety of models have been proposed for predicting nucleosome occupancy that are classified into categories of bioinformatics 19 39 40 41 42 43 44 45 and energetics of nucleosomal DNA 46 47 48 49 50 51 52 53 54 . Bioinformatics models learn various sequence features, such as dinucleotide distributions and oligonucleotide motif frequency from a large quantity of nucleosomes 39 40 41 42 43 .…”

“…There are a number of energetics models designed to predict nucleosome formation energy, nucleosome occupancy and positions 46 47 48 49 50 51 52 53 54 . A model 51 that took into account the deformations of DNA helical twist, roll and tilt achieved a moderate correlation between its prediction and experimental nucleosome occupancy (R = 0.45, P < 0.0001, on yeast chromosome III).…”