Pheg

Pheg predicts human essential genes from nucleotide composition and association information using the λ-interval Z curve to enable sequence-based identification of essential genes.

Key Features:

• λ-interval Z curve representation: Represents nucleotide composition and association information by applying the λ-interval Z curve method to DNA sequences.
• Support Vector Machine (SVM) classifier: Uses an SVM model to classify genes as essential or non-essential based on the extracted sequence features.
• Performance: Achieves an Area Under the Curve (AUC) greater than 0.88 in both 5-fold cross-validation and jackknife tests.
• Prediction of additional essential genes: Identified additional essential genes that were overlooked in experimental datasets.
• Homology insights: Among newly predicted essential genes, 20 were homologous to known mouse essential genes.

Scientific Applications:

• Gene function analysis: Facilitates identification of genes critical for cell survival and core biological processes.
• Cancer research: Supports prediction of gene essentiality in human cancer cell lines to inform potential therapeutic targets.
• Comparative genomics: Enables cross-species comparisons of essential genes and assessment of evolutionary conservation.

Methodology:

Extracts nucleotide composition and association features using the λ-interval Z curve, classifies genes with a Support Vector Machine, and evaluates performance using 5-fold cross-validation and jackknife tests.