CRH

CRH resolves domain matches onto stretches of query proteins by selecting an optimal subset of candidate matches with minimal or no overlap, including handling discontinuous domains.

Key Features:

• Dynamic-programming algorithm: Implements a dynamic-programming algorithm in open-source C++ to assign optimal, non-overlapping domain matches, including discontinuous domains.
• High-performance throughput: Processes approximately 1 million hits per second while maintaining reasonable memory usage.
• Large-dataset optimization: Engineered to reduce CPU time and memory requirements for very large datasets compared to existing algorithms.
• Input and output formats: Accepts multiple input formats and produces outputs in plain text, JSON, and graphical HTML.
• Benchmarking and validation: Benchmarking against an existing algorithm yields comparable results with slight improvements in accuracy.

Scientific Applications:

• Protein structure prediction: Improves domain assignments used in protein structure prediction.
• Functional annotation: Supports functional annotation by providing precise domain mappings.
• Evolutionary studies: Enables evolutionary analyses through resolved domain architectures across homologs.
• Structural biology: Assists structural biology research by clarifying domain boundaries and discontinuities.

Methodology:

Implements a dynamic-programming algorithm in C++, assigns non-overlapping subsets of domain matches including discontinuous domains, supports multiple input formats with outputs in plain text, JSON, and graphical HTML, and was benchmarked against an existing algorithm achieving ~1 million hits per second with reasonable memory usage.