Unified cross-modality integration and analysis of T-cell receptors and T-cell transcriptomes

Single-cell RNA sequencing and T-cell receptor sequencing (scRNA-seq and TCR-seq, respectively) technologies have emerged as powerful tools for investigating T-cell heterogeneity. However, the integrated analysis of gene expression profiles and TCR sequences remains a computational challenge. Herein, we present UniTCR, a unified framework designed for the cross-modality integration and analysis of TCRs and T-cell transcriptomes for a series of challenging tasks in computational immunology. By utilizing a dual-modality contrastive learning module and a single-modality preservation module to effectively embed each modality into a common latent space, UniTCR demonstrates versatility across various tasks, including single-modality analysis, modality gap analysis, epitope-TCR binding prediction and TCR profile cross-modality generation. Extensive evaluations conducted on multiple scRNA-seq/TCR-seq paired datasets showed the superior performance of UniTCR. Collectively, UniTCR is presented as a unified and extendable framework to tackle diverse T-cell-related downstream applications for exploring T-cell heterogeneity and enhancing the understanding of the diversity and complexity of the immune system. ### Competing Interest Statement The authors have declared no competing interest.