CATT

CATT detects and characterizes T-cell receptor (TCR) complementarity-determining region 3 (CDR3) sequences from TCR-seq, RNA-Seq, and single-cell TCR(RNA)-Seq data for ultra-sensitive profiling of TCR repertoires.

Key Features:

• Input data support: Processes TCR-seq, RNA-Seq, single-cell TCR(RNA)-Seq, and other sequencing data containing TCR information.
• de Bruijn micro-assembly: Uses a de Bruijn graph-based micro-assembly algorithm to reconstruct CDR3 sequences.
• Error correction: Implements a data-driven error correction model to reduce sequencing and amplification errors.
• Bayesian inference: Applies Bayesian inference algorithms for sequence inference and confidence estimation.
• Self-adaptive sensitivity: Self-adaptively and ultra-sensitively characterizes CDR3 repertoires to improve detection of low-abundance clonotypes.
• Short-read and small-sample performance: Optimized for datasets with short read lengths, small sizes, and single-cell sequencing.
• Benchmarking: Demonstrated improved recall and precision relative to current methods on in silico and experimental datasets.

Scientific Applications:

• CDR3 repertoire profiling: High-sensitivity detection and quantification of TCR CDR3 repertoires from bulk and single-cell sequencing.
• TCR function studies: Facilitates investigation of TCR roles in antigen recognition and T-cell immunology.
• Cancer immunology: Enables analysis of TCR repertoires relevant to cancer immunology research.

Methodology:

Computational methods explicitly include a de Bruijn graph-based micro-assembly algorithm, a data-driven error correction model, and Bayesian inference algorithms for self-adaptive CDR3 characterization.