PhIP-seq enables AIH disease prediction and identifies both known and novel highly specific autoantibody targets. (a) Phage-display library design, and an overview of methods to apply PhIP-seq to evaluate the AIH and control cohort. (b) Summary of the customized bioinformatics analysis pipeline applied to analyze next-generation sequencing output of PhIP-seq data. (c) ROC curve analysis for prediction of AIH versus all control patients, noting AUC. (d) Heatmap of significantly enriched peptides, which are highly specific to AIH (>99% specific). Depicted in the heap map are top hits where multiple overlapping peptides were immunoprecipitated by PhIP-seq (peptide name at bottom, right axis). The top legend indicates whether samples correspond to case patients (AIH: pink) or controls (healthy controls: blue; PBC: yellow; MASLD: green; RA: purple); boxes are shaded by Z-score of enrichment (legend bottom left), and Z <3 is shaded gray. Highlighted on the right axis, in bold, are the hits with the highest mean Z-scores across all enriched peptides. The panel below zooms in on results of Z-score of enrichment for the AIH patients positive for any of the top targets in bold, with a top legend indicating their AIH treatment regimen (red: on treatment without steroids; blue: on treatment containing steroids; yellow: no reported treatment). ROC curve, receiver operating characteristic curve.