NuPoP

NuPoP predicts nucleosome positioning and occupancy from genomic DNA sequences using a duration Hidden Markov Model to inform analyses of chromatin organization and gene regulation.

Key Features:

• Duration Hidden Markov Model: Employs a duration HMM that models nucleosome states and explicitly represents both the nucleosome core particle and linker DNA.
• Linker length distribution modeling: Explicitly models variation in linker DNA length to improve positional accuracy across genomic regions.
• Base composition re-scaling: Applies base composition re-scaling when transferring models trained on yeast to other species to compensate for sequence composition differences.
• Performance metrics from simulations: Simulation studies report increased sensitivity and reduced false discovery rate compared with existing methods due to linker modeling and re-scaling.

Scientific Applications:

• Genome organization and chromatin structure: Generates nucleosome occupancy and positioning maps to analyze chromatin organization across genomes.
• Gene regulation and DNA accessibility: Supports investigation of how nucleosome positioning influences gene regulation and DNA accessibility.
• Comparative genomics: Enables cross-species comparison of nucleosome organization via model transfer and base composition adjustment.

Methodology:

Uses a duration Hidden Markov Model to represent nucleosome core particle and linker DNA states, models linker length distributions, and applies base composition re-scaling when adapting yeast-trained models to other organisms.