The case for caution in the application of whole-exome sequencing data for immune repertoire analysis

The analysis of the adaptive immune receptor repertoire is critical in disease research. While specialized immune repertoire sequencing (IR-seq) is the dedicated method for receptor repertoire profiling, the abundance of existing WES data has led to its frequent repurposing for repertoire analysis; however, its reliability for this application remains poorly evaluated. To address this, we first assessed the VDJ gene coverage by commercial WES probes. Next, using a multi-omics dataset (matched IR-seq, RNA-seq, WES), we compared repertoire inferences derived from WES and RNA-seq results against IR-seq. This comparison was followed by an analysis correlating WES probe matching efficiency with the failure to identify V/J gene segments. We found that WES probes miss a substantial fraction of individual V, D, and J genes and provide uneven coverage across these genes. WES consistently showed significantly lower performance than RNA-seq, detecting fewer clonotypes and yielding greater abundance deviation. The underestimation of clonotype richness in WES was confirmed through external validation using multiple public datasets. Our results illustrate the inherent limitations of WES for repertoire profiling, serving as a cautionary note for the appropriate interpretation of exome-derived repertoire information in research.