AVARDA

AVARDA deconvolves antibody-binding signals in VirScan programmable phage-display datasets to identify virus-specific and cross-reactive antibodies across the human virome using sequence alignment and probabilistic modeling.

Key Features:

• Multi-module architecture: Modular analysis components operate on VirScan peptide-level data.
• Probabilistic assessment: Employs probabilistic models to infer infections at species-level resolution.
• Sequence alignment: Aligns all peptides in the VirScan library against each other and against known human viral sequences.
• Cross-reactivity modeling: Accounts for antibody cross-reactivity among related viruses to improve specificity.
• Library representation correction: Adjusts for disproportionate representation of individual viruses within the phage-display library.

Scientific Applications:

• Diagnosis of encephalitis: Analyzed acute and convalescent sera to confirm or detect responses to human herpesviruses 1, 3, 4, 5, and 6, increasing diagnostic yield (reported +44%).
• Association studies: Applied to investigate links between viral infections and chronic diseases, implicating enterovirus in type 1 diabetes and herpesvirus in lupus.

Methodology:

Analyzes VirScan peptide-level data by aligning library peptides to each other and to known human viral sequences, then applies probabilistic models that account for antibody cross-reactivity and uneven library representation to infer species-level infections.