PHONI

PHONI computes incremental matching statistics for patterns within highly repetitive and compressed genomic databases to support large-scale sequence comparison.

Key Features:

• Memory Efficiency: Uses the MONI index to preprocess the input text and reduce working memory requirements on large genomic datasets.
• Streaming Model: Implements a streaming computation strategy based on the approaches of Bannai et al. and Rossi et al., trading some execution speed for a streaming architecture.
• Parallel Processing: Supports parallel computation of matching statistics for multiple long patterns, such as entire human chromosomes, while limiting RAM usage.
• Online Low-Latency Computation: Enables online detection when a pattern becomes incompressible relative to the database, supporting low-latency analysis.

Scientific Applications:

• Large-scale sequence comparison: Computing matching statistics for long genomic patterns against large, repetitive genomic databases.
• Analysis of repetitive genomes: Detecting incompressible patterns and comparing sequences in highly repetitive or compressed genomic collections where memory or computational efficiency is critical.

Methodology:

Preprocesses the input text using the MONI index and then applies a heuristic incremental matching-statistics computation that supports both batch processing of multiple patterns and online low-latency recognition.