PDGA

PDGA generates peptide sequences to explore and populate chemical space with peptide analogs similar to target molecules for drug discovery using a genetic algorithm guided by the MXFP (Molecular eXtended Fingerprint).

Key Features:

• Chemical Space Exploration: Navigates a defined chemical space organized by molecular structures and properties using the MXFP atom-pair fingerprint that encodes molecular shape and pharmacophoric features.
• Diverse Peptide Topologies: Produces peptide sequences with linear, cyclic, polycyclic, and dendritic chain topologies.
• High-Similarity Analog Generation: Generates high-similarity analogs to bioactive peptides and non-peptide targets, including known active analogs.
• Fitness Evaluation by MXFP: Optimizes peptide candidates using fitness criteria derived from MXFP measures of molecular shape and pharmacophoric similarity.
• Interactive MXFP Visualization: Provides a 3D-map of the MXFP property space accessible at http://faerun.gdb.tools/ for inspection of generated-peptide distributions.

Scientific Applications:

• Drug Discovery and Lead Optimization: Design and optimization of peptide candidates closely related to specified target molecules.
• Peptide Mimicry of Non-Peptide Targets: Create peptide analogs that replicate structural and pharmacophoric features of non-peptide bioactive molecules.
• Exploration of Novel Chemical Entities: Expand accessible chemical space to identify new peptide-based bioactive molecules with desired activities.

Methodology:

PDGA implements a genetic algorithm that iteratively evolves peptide sequences by simulating natural selection and optimizes peptides toward higher similarity to a target using predefined fitness criteria derived from MXFP's measures of molecular shape and pharmacophoric features.