Chromatin states (30 state model, BG3 cells)

To derive a nine-state joint chromatin state model for the S2 and BG3 cells, the genome was first divided into 200 bp bins, and the average enrichment level was calculated within each bin based on unsmoothed M values (using all histone enrichment profiles and Pc to discount the genome-wide difference in S2 H3K27me3 profiles). The bin-average values of each mark were shifted by the genome-wide mean, scaled by the genome-wide variance, and quantile-normalized between the two cells. An HMM model with multivariate normal emission distributions was generated using the data from both cell lines (30 seeding configurations determined with K-means clustering were used), and the Baum-Welch algorithm. States with minor intensity variations (Euclidian distance of mean emission values < 0.15) were merged. Larger models (up to 30 states) were examined, and the final number of states was chosen for optimal interpretability.