[Identification of nucleosome positioning using support vector machine method based on comprehensive DNA sequence...

Area of Science:

• Computational Biology
• Genomics
• Bioinformatics

Background:

• Nucleosome positioning is crucial for gene regulation.
• Accurate prediction of nucleosome occupancy is essential for understanding genomic processes.
• Existing methods for nucleosome positioning prediction have limitations.

Purpose of the Study:

• To develop a novel computational model for predicting nucleosome positioning.
• To enhance the accuracy and reliability of nucleosome identification in DNA sequences.
• To evaluate the model's performance across different species.

Main Methods:

• Construction of a nucleosome feature model (CSeqFM) using z-curve theory and position weight matrices (PWM).
• Transformation of nucleosome sequence datasets into 3D coordinates and calculation of PWM.
• Integration of features and calculation of Euclidean distance for SVM training and ten-fold cross-validation.

Main Results:

• High accuracy in identifying nucleosome positioning for *S. cerevisiae* (97.1% sensitivity, 96.9% specificity, 94.2% accuracy, 0.89 MCC, 0.980 AUC).
• Superior performance compared to existing z-curve methods.
• Effective and stable nucleosome positioning identification in *C. elegans*, *H. sapiens*, and *D. melanogaster* (AUCs > 0.90).

Conclusions:

• The CSeqFM model demonstrates strong reliability and excellent identification performance for nucleosome positioning.
• CSeqFM outperforms other methods like iNuc-STNC and iNuc-PseKNC.
• The model shows broad applicability and stability across diverse species.