IMPre

IMPre predicts germline Variable (V), Diversity (D), and Joining (J) genes and alleles from deep-sequencing T-cell receptor (TCR) and B-cell receptor (BCR) repertoires.

Key Features:

• Seed_Clust algorithm: Implements the "Seed_Clust" clustering algorithm to assemble a multiway tree of sequences optimized for V(D)J rearrangement characteristics.
• Training and benchmarking: Trained on human TRB and IGH samples with reported accuracies of 97.7% for TRBV, 100% for TRBJ, 92.9% for IGHV, and 100% for IGHJ.
• Robustness and subsampling: Demonstrates stable performance across varying data quantities and subsampling experiments.
• Cross-species and sequence-length validation: Validated on rhesus monkey datasets and human long sequences, maintaining high accuracy and stability.
• Novel gene and allele discovery: Designed to leverage accumulating high-throughput sequencing of TCR and BCR repertoires to identify novel germline genes and alleles.

Scientific Applications:

• Adaptive immune response studies: Enables characterization of V(D)J gene and allele repertoires to study adaptive immune responses.
• Comparative immunogenomics: Supports comparative analyses of germline V(D)J repertoires across species such as human and rhesus monkey.
• Novel allele discovery: Facilitates identification of previously uncharacterized V, D, and J alleles from deep-sequencing repertoire data.

Methodology:

Integrates deep-sequencing repertoire data with clustering algorithms; the Seed_Clust algorithm clusters sequences into a multiway tree structure optimized for V(D)J rearrangement characteristics.