ClareV: a contrastive learning framework for context-aware TRBV representations in TCR repertoires.

Background: V gene segments of the T cell receptor (TCR) Methods: We introduce ClareV, a contrastive learning framework that partitions T cell receptor repertoires into V gene-specific groups to extract data-driven embeddings capturing these dynamics. We evaluated ClareV across three TCR repertoire cohorts spanning cytomegalovirus (CMV), gastric cancer, and multi-cancer settings. Results: In the Emerson2017 CMV cohort, the ClareV Random Forest (RF) Fusion model improved AUC by 14.3% over the strongest V-family usage-only baseline (linear SVM) and by 13.3% over the strongest V-gene usage-only baseline (Random Forest). These results indicate that ClareV embeddings capture additional repertoire-level information beyond V-frequency summaries. Ablation studies showed that frequency-weighted bagging and contrastive V-bag learning both contributed to performance, outperforming simple pooling and non-contrastive learned-bag controls. In the CMV cohort, downstream embedding analyses showed that ClareV representations partially recapitulated IMGT-defined V-gene structure while retaining a CMV-responsive, non-sequence component that was largely orthogonal to classical V-usage signals. Discussion: These findings support adaptive V-gene representations as a compact feature space that complements V-frequency summaries for repertoire-level classification.