Fec

Fec corrects errors in long reads from third-generation sequencing technologies (PacBio and nanopore) by using a two-round overlapping strategy together with a base-level alignment caching mechanism to accelerate and extend overlap-based error correction.

Key Features:

• Two-Round Overlapping Strategy (round 1): Uses a large window size of 20 in the first round to rapidly identify overlaps and correct the majority of read errors.
• Adaptive Second Round: Executes a second round with a smaller window size of 5 for reads with insufficient overlaps from the first round to find additional overlaps and correct remaining errors.
• Caching Mechanism: Stores base-level alignment information in a cache, checks the cache before performing alignments, and updates the cache when new alignments are computed to avoid redundant computation.
• Alignment Reuse: Reuses cached alignment information to deduce corrections without recomputing alignments for previously processed overlaps.
• Performance: Empirical tests on nine datasets report speed-ups of 1.24–38.56× versus MECAT, CANU, and MINICNS on five PacBio datasets and 1.16–27.8× versus NECAT and CANU on four nanopore datasets.
• Targeted Read Types: Specifically developed for long reads generated by PacBio and nanopore sequencing technologies.

Scientific Applications:

• Assembly pipelines: Can be integrated as a component within genome assembly pipelines to improve corrected-read quality prior to assembly.
• High-coverage sequencing projects: Applied to high-coverage PacBio and nanopore datasets to reduce computational burden of overlap-based error correction.

Methodology:

Fec applies a two-round overlapping approach (window sizes 20 then 5 for insufficient overlaps), stores and checks base-level alignment information in a cache, reuses cached alignments to deduce corrections, and updates the cache when alignments are performed.