Supplementary Table 6 - Occurrence of contiguous runs of unpreferred codons in relation to linker sequence Length of run (X) in codons Total number of runs (N) Number of runs in which at least one nucleotide is called as linker (F) E>O E=3 unpreferred codons from 5788 mapped S. cerevisiae genes were analyzed with regard to the nucleosome occupancy calls of their nucleotides. Unpreferred codons, defined as codons exhibiting uneven usage between highly and lowly expressed genes in the S. cerevisiae genome, are as follows (taken from Neafsey and Galagan (2007)):AAA AAT ACA ACG AGC AGG AGT ATA CAG CAT CCC CCG CCT CGA CGC CGG CTA CTC CTG CTT GAG GAT GCA GCG GGA GGC GGG GTA GTG TAT TCA TCG TGC TTA TTT. Runs of length X are observed N times in the data. Run length membership is mutually exclusive, i.e. runs of 4 codons do not also contribute to the count of runs of 3 codons etc. For N run of length X, F runs contribute at least partially to linker sequence."Runs with >=100 instances were analyzed. To test whether runs of unpreferred codons were enriched in linker we sampled random runs of length X from random genes in our sample N times and, as before, determined the number of runs with at least one nucleotide called as linker. We repeated this procedure 100 times and asked how often we observe more than/less than F runs that participate in linker. For example, for runs of 8 codons, we expect to find 3362 randomly chosen runs to contain less (E=6 unpreferred codons to be more common in linker than expected. Why we observe the reverse tendency for runs <6 is unclear. A likely part of explanation might be that highly expressed genes are depleted for long codon runs while also mostly composed of fuzzy sequence so that clear-cut linker sequence is rarely called. The exact nature of the relationship between runs of unpreferred codon and linker remains to be explored in greater detail in the future.