Figure S1 - Preferentially Quantized Linker DNA Lengths in <em>Saccharomyces cerevisiae</em>

Linker DNA length distribution (FL(d)) for random (A) or shuffled (B) dinucleosome DNA sequences, using the DHMM-kernel smoothing method. (A) Random dinucleosome sequences simulated based on the product multinomial model with the same base composition and the same length distribution as in the real data. The DHMM-kernel procedure was followed exactly as in Methods. In both simulations, the resulting linker length distribution varied between trials, and the ∼10 bp periodicity disappeared in general. (B) As in (A) except using sequences obtained by shuffling true dinucleosome sequences while keeping the dinucleotide frequency fixed. (0.41 MB EPS)