Single-cell transcriptome and antibody repertoire sequencing of antigen-specific and non-specific B cells

The field of antibody discovery relies heavily on extensive experimental screening of B cells from immunized animals. Machine learning (ML)-guided prediction of antigen-specific B cells would offer the potential to accelerate this process. However, training such ML models would require sufficient training data with antigen-specificity labeling. Here, we introduce a dataset of single-cell transcriptome and antibody repertoire sequencing of B cells from immunized mice. Importantly, B cells are labeled as antigen-specific or non-specific through experimental selections. We identify gene expression patterns associated with antigen-specificity and assess their antibody sequence diversity. Subsequently, we benchmark various ML classification models, including linear, non-linear models, which are trained on different feature combinations of gene expression and antibody repertoires. Additionally, we assess transfer learning approaches using features from general and antibody-specific protein language models (PLMs). Our findings show that gene expression-based models outperform sequence-based models for antigen-specificity predictions and may contribute to computational-guided discovery of antibodies.