plifepred

plifepred predicts and assists the design of therapeutic peptide half-lives in mammalian blood using computational models derived from experimentally characterized natural and modified peptides.

Key Features:

• Dataset composition: Models were developed from a dataset of 261 peptides, comprising 163 natural and 98 modified variants with experimentally determined half-lives.
• Descriptor types: Models use chemical descriptors, including amino acid composition and PaDEL descriptors.
• PaDEL descriptor models: A model using 43 PaDEL descriptors achieved a correlation coefficient of 0.692 between predicted and actual half-lives.
• Natural-peptide models: Models using amino acid composition on 163 natural peptides achieved a maximum correlation of 0.643, and a model using 45 PaDEL descriptors on natural peptides achieved a correlation of 0.743.
• Peptide half-life design: Enables computational design strategies to tailor peptide half-lives for therapeutic applications.

Scientific Applications:

• Therapeutic peptide design: Guiding design of peptides with tailored half-lives for therapeutic development.
• Pharmacokinetic estimation: Estimating peptide half-life in mammalian blood for pharmacokinetic assessment.
• Comparative analysis: Comparing half-life effects between natural and chemically modified peptides.

Methodology:

Models were constructed from a 261-peptide dataset using chemical descriptors (amino acid composition and PaDEL descriptors, including 43- and 45-descriptor sets) to predict half-life in mammalian blood, with reported correlation coefficients of 0.692, 0.643, and 0.743.