RoboCOP: jointly computing chromatin occupancy profiles for numerous factors from chromatin accessibility data

Area of Science:

• Molecular Biology
• Genomics
• Computational Biology

Background:

• Chromatin structure, composed of DNA and proteins, regulates gene expression.
• Understanding transcription factor (TF) and nucleosome binding is key to gene regulation.
• Current methods like antibody-based assays or chromatin accessibility assays have limitations in identifying multiple factors simultaneously.

Purpose of the Study:

• To develop a novel computational model, RoboCOP, for predicting genome-wide nucleosome and TF occupancy.
• To integrate chromatin accessibility data with nucleotide sequence information for enhanced prediction accuracy.
• To elucidate the protein-binding landscape and chromatin dynamics in yeast, including under stress conditions.

Main Methods:

• Developed RoboCOP, a multivariate state space model.
• Integrated chromatin accessibility data (MNase-seq, ATAC-seq) with nucleotide sequence information.
• Applied the model to the yeast genome to predict occupancy of nucleosomes and 150 TFs.

Main Results:

• RoboCOP accurately computes genome-wide probabilistic scores for nucleosome and TF occupancy.
• The model demonstrates superior predictive performance compared to existing methods.
• A detailed chromatin occupancy profile of the yeast genome under cadmium stress was generated, revealing dynamic changes.

Conclusions:

• RoboCOP offers a powerful approach to jointly infer nucleosome and TF occupancy from chromatin accessibility data and sequence.
• The model significantly enhances our ability to study genome-wide protein-DNA interactions and gene regulation.
• This work provides new insights into chromatin dynamics and transcriptional regulation in response to environmental stress.