CodonAdjust

CodonAdjust optimizes codon nucleotide composition to produce codon mixtures that match specified amino acid (AA) profiles for mutagenesis library design.

Key Features:

• AA profile matching: Generates codon mixtures that mimic specified amino acid (AA) composition targets.
• Optimization formulation: Formulates codon design as a non-convex mathematical optimization problem minimizing the squared error between targeted and computed AA profiles.
• Mathematical optimization: Employs mathematical optimization techniques to solve the non-convex design problem.
• Automated weight tuning: Automates tuning of amino-acid weights within objective functions to replace manual parameter adjustment.
• Configurable options: Provides four distinct options for customization of objective functions and constraint settings.
• Threshold frequency constraints: Allows specification of threshold frequencies for amino acids as constraints in the design.
• Library generation support: Produces codon designs intended for construction of mutagenesis libraries with targeted AA compositions.

Scientific Applications:

• Mutagenesis library design: Design of mutagenesis libraries that reflect precise amino-acid composition targets.
• Antibody CDRs: Applied to complementarity-determining regions (CDRs) of antibodies in experimental testing.
• T-cell receptors (TCRs): Applied to T-cell receptor sequence regions in experimental testing.
• Protein family analysis: Validated on large-scale datasets including millions of AA profiles from Pfam to assess robustness across protein families.

Methodology:

Formulates the design task as a non-convex optimization problem solved with mathematical optimization techniques to minimize the squared error between target and computed AA profiles, includes automated tuning of amino-acid weights within the objective, and supports four configurable options including amino-acid frequency threshold constraints.