Fig. 1

Comparison of prediction accuracy between chemical map-based and MNase-seq-based models for yeasts. A Distance between predicted nucleosomes and nearest in vivo nucleosomes. Fraction of predicted nucleosomes on the Viterbi path located at a particular distance from the nearest in vivo nucleosomes is plotted. Species of the target genome and for model construction were both S. cerevisiae. Magnified view is presented inside each plot. Arrows point to the fractions of predicted nucleosomes that are 1-bp apart from in vivo nucleosomes. B Matching rate between predicted and in vivo nucleosomes with variable matching windows. The species of genomic sequences used for prediction (Target) and the species of the chemical map used for model construction (Model) are indicated. “Sc” stands for S. cerevisiae, whereas “Sp” stands for S. pombe. Indicated R packages were used for predictions. C Matching of nucleosomes on the Viterbi paths with in vivo unique and redundant nucleosomes at 2-bp resolution. Note that the redundant nucleosome datasets contain both unique nucleosomes (red), which are representative, and non-representative nucleosomes (orange)