RADProc

RADProc performs de novo locus assembly from Restriction site-associated DNA sequencing (RADseq) data to identify loci and assign sequence reads for genotyping in population studies.

Key Features:

• Graph-Based Data Structure: Represents all sequence reads and their similarity relationships in a graph to store sequence-comparison results and avoid redundant similarity calculations.
• Precomputed Sequence Similarities: Precomputes and stores sequence-comparison results within the graph to accelerate downstream analyses.
• Efficient Parameter Sweeping: Uses the precomputed graph to enable rapid de novo locus formation across multiple parameter settings for fast parameter exploration.
• Clustering-Based Acceleration: Incorporates a clustering approach to speed nucleotide-distance calculations and improve processing efficiency on large datasets.
• Performance Benchmarks: Demonstrated reduced runtimes relative to Stacks (e.g., 20 Carcinus maenas samples in 2 h 40 min versus 78 h with Stacks; 16 Salmo trutta L. samples in 23 h versus 263 h with Stacks).

Scientific Applications:

• Population genomics: Enables genotyping and locus discovery from RADseq data for population-level studies.
• Empirical dataset analysis and benchmarking: Applied to green-crab (Carcinus maenas) and brown trout (Salmo trutta L.) samples for performance and accuracy comparisons.
• Parameter optimization: Facilitates rapid parameter sweeps to identify assembly settings that impact locus formation and downstream analyses.

Methodology:

Constructs a graph of sequence reads and similarity relationships, stores precomputed sequence-comparison results in the graph, and applies clustering to accelerate nucleotide-distance calculations for de novo locus formation across parameter settings.