K-PAM

K-PAM predicts and models Klebsiella K- and O-antigen serotypes and surface antigen 3D structures to support serotyping, hypervirulence marker detection, and structural analyses.

Key Features:

• Serotype Prediction: K-PAM predicts Klebsiella K-types by analyzing proteins involved in capsular polysaccharide (CPS) biosynthesis and surface expression, assigning reliability scores (ARS) with values: Wzx 98.5%, Wzy 97.5%, WbaP 97.2%, Wzc 96.4%, Wzb 94.3%, WcaJ 93.8%, Wza 79.9%, and Wzi 37.1% (improved to 80.8% using a fragment-based approach).
• Improved Accuracy: Use of multiple genes/proteins enhances accuracy, reduces serotype multiplicity, and distinguishes identical K-locus types such as KN3 and K35.
• Ancestral Serotype Identification: Identifies ancestral serotypes of Klebsiella species to aid outbreak surveillance and evolutionary analysis.
• O-type Prediction: Predicts O-types using Wzm and Wzt, each with an ARS of 85.7%.
• Hypervirulent Strain Identification: Detects hypervirulent strains by identifying marker genes rmpA, rmpA2, iucABCD, iroBCDN, and aid.
• 3D Structure Modeling: Provides a repository of modeled 3D structures comprising 11 O-antigens and 79 K-antigens for structural and antigenic analyses.

Scientific Applications:

• Outbreak Surveillance: Supports outbreak surveillance and epidemiological tracking by providing serotype and ancestral-serotype information.
• Hypervirulence Differentiation: Enables differentiation of hypervirulent and non-hypervirulent Klebsiella strains via marker gene detection.
• Structural Biology: Facilitates structural biology and antigenicity studies using modeled 3D structures of K- and O-antigens.
• Evolutionary Analysis: Enables evolutionary analysis of serotype variation through comparative serotype assignments and ancestral-serotype identification.

Methodology:

K-PAM integrates protein sequence data with reliability scoring to predict serotypes, applies a fragment-based enhancement for Wzi to improve accuracy, employs automated testing and validation of predicted serotypes, and integrates multiple gene/protein analyses to refine predictions and reduce ambiguity.