nucleoSim

nucleoSim simulates synthetic nucleosome-mapped next-generation sequencing (NGS) reads to model nucleosome positions for analyses of chromatin structure and gene regulation.

Key Features:

• Synthetic NGS read generation: Generates forward and reverse reads that cover nucleosome regions to produce synthetic sequencing datasets.
• Read positioning distributions: Samples read positions using Normal, Student's t-distribution, or Uniform distributions.
• Bayesian statistical model: Implements a Multinomial-Dirichlet classification combined with hierarchical mixture distributions.
• Reversible-jump MCMC inference: Estimates the number and positions of nucleosomes using reversible jump Markov chain Monte Carlo.
• Evaluation and comparison: Performance assessed on simulated data and MNase-Seq datasets from Saccharomyces cerevisiae and compared to PING and NOrMAL.
• Relation to ChIP mapping approaches: Aligns with concepts from ChIP-chip and ChIP-Seq genomic mapping techniques.
• Alternative to EM and heuristics: Addresses limitations of Expectation-Maximization (EM) algorithms and heuristic strategies for estimating nucleosome numbers.

Scientific Applications:

• Chromatin structure analysis: Simulate nucleosome maps to investigate chromatin organization and nucleosome positioning.
• Gene regulation studies: Model how nucleosome placement affects transcription factor access and gene expression regulation.
• Algorithm benchmarking: Provide synthetic datasets for benchmarking nucleosome-calling methods such as PING and NOrMAL.
• Yeast nucleosome studies: Apply to analysis and validation with MNase-Seq datasets from Saccharomyces cerevisiae.

Methodology:

Computational methods include simulation of forward and reverse NGS reads with read positions sampled from Normal, Student's t, or Uniform distributions; a Bayesian Multinomial-Dirichlet classification with hierarchical mixture distributions; parameter and model inference via reversible-jump MCMC; and evaluation against simulated data and MNase-Seq from Saccharomyces cerevisiae with comparisons to PING and NOrMAL.