SLAPNAP

SLAPNAP predicts HIV-1 Envelope pseudovirus neutralization sensitivity and resistance to broadly neutralizing antibodies (bnAbs) by training machine-learning models on sequence and CATNAP data to inform bnAb regimen selection for prevention studies.

Key Features:

• Data Integration: SLAPNAP leverages the Compile, Analyze and Tally NAb Panels (CATNAP) database to compile HIV Envelope sequences and associated neutralization measurements for analysis.
• Model Development: The tool develops predictive models using machine-learning algorithms, including cross-validation-based ensemble predictors, trained on amino acid sequence features of the HIV Envelope.
• Variable Importance Analysis: SLAPNAP quantifies variable importance of sequence features to identify specific amino acid positions and motifs that influence neutralization sensitivity and resistance.
• Regimen Ranking: SLAPNAP predicts in vitro neutralization resistance and ranks single bnAb regimens and bnAb combinations by predicted potency and breadth.

Scientific Applications:

• Evaluation of bnAb regimens: Predicting neutralization outcomes to support selection and down-selection of bnAb regimens for prevention studies.
• Candidate identification: Prioritizing promising bnAb candidates and combinations based on predicted neutralization performance.
• Mechanistic insight: Elucidating molecular determinants of resistance by linking amino acid sequence features to neutralization sensitivity.
• Trial design support: Informing selection of bnAb combinations for randomized prevention trials by comparing predicted potency and breadth.

Methodology:

SLAPNAP compiles analytic datasets from the CATNAP database, trains cross-validation-based ensemble machine-learning predictors on amino acid sequence features of the HIV Envelope, computes variable importance measures, and summarizes predicted in vitro neutralization resistance.