Echidna

Echidna simulates single-cell immune receptor repertoires and matched transcriptomes as an R package to produce ground-truth datasets for analysis and benchmarking of clonal selection, expansion, and immune evolution.

Key Features:

• Integrated receptor and transcriptome simulation: Generates integrated single-cell immune receptor repertoires for B cell receptor (BCR) and T cell receptor (TCR) sequences with matched gene expression profiles.
• User-tunable parameters: Allows adjustment of clonal expansion, germline gene usage, somatic hypermutation, and transcriptional phenotypes.
• Time-resolved B cell evolution simulation: Simulates evolutionary trajectories that produce mutational networks with selection histories, including class-switching events and B cell subtype annotations.
• Benchmarking and validation support: Enables comparison of known simulated clonal lineages and networks against those inferred from BCR sequences to evaluate clonotyping, phylogenetics, transcriptomics, and machine learning methods.
• Experimental data integration: Provides a framework for incorporating experimental data into simulations.
• Annotated single-cell sequencing output: Produces annotated single-cell sequencing data with explicit control over biological features.

Scientific Applications:

• Method benchmarking: Benchmark clonotyping, phylogenetics, transcriptomics, and machine learning approaches using simulated ground-truth repertoires and expression data.
• Clonal selection and expansion studies: Investigate dynamics of clonal selection and expansion under controlled simulation parameters.
• B cell evolution and class-switching analysis: Study mutational networks, selection histories, and class-switch recombination in simulated B cell lineages.
• Software development and validation: Support development and validation of bioinformatic tools that analyze single-cell immune receptor and transcriptome data.

Methodology:

Implemented as an R package that simulates single-cell immune receptor repertoires and matched transcriptomes, generates time-resolved mutational networks, models somatic hypermutation and class-switching, and allows incorporation of experimental data into simulations.