ASAP-SML

ASAP-SML identifies distinguishing sequence and structural features in antibody repertoires by comparing targeting antibody sets to reference non-targeting sets using statistical testing and machine learning.

Key Features:

• Feature Extraction: Extracts feature fingerprints from antibody sequences including germline gene usage, CDR canonical structures, isoelectric points, and frequent positional motifs.
• Machine Learning Integration: Applies machine learning algorithms to feature fingerprints to detect patterns and combinations that differentiate targeting from non-targeting antibodies.
• Statistical Significance Testing: Performs statistical testing on identified features to validate their relevance and reduce spurious associations.
• Output and Recommendations: Produces reports highlighting key distinguishing features and recommends salient sequence features for designing novel antibodies.

Scientific Applications:

• Matrix metalloproteinase (MMP) analysis: Applied to antibody sequences targeting matrix metalloproteinases (MMPs), a family of zinc-dependent enzymes implicated in cancer progression and pathological inflammation, to identify features distinct from non-targeting datasets.
• Therapeutic antibody design: Recommends specific sequence and structural features to guide design of novel antibodies targeting similar proteins for therapeutic development.

Methodology:

Inputs comprise a set of antibody sequences known to bind or inhibit specific targets alongside a reference non-targeting dataset; the pipeline extracts features to create per-sequence fingerprints; machine learning analyzes fingerprints to identify distinguishing features which are then subjected to statistical significance testing; outputs include reports and design recommendations.