Fig. 1

An overview of NucHMM workflow. A The initialization step combines several existing tools to construct nucleosome-level HMM training sequences. B The training step includes a selection of the best model and two rounds of HMM training for the best selected model. BW algorithm is applied to acquire the transition probability matrix and the mark-state matrix derived from the emission probability matrix. C The functioning step performs functional screening on the nucleosomes based upon the genomic location, nucleosome array number, nucleosome spacing, phasing, and positioning and finally identifies the functional nucleosome states