MS release

1.1.1

release

Robust error model and more features

New features

• Minor based error model now adjusts for false discovery rate (sequencing errors that are considered PCR errors, multiple testing correction) and limited recall due to stochastic sampling of PCR minors. Poisson distribution is used to model both sequencing error rate and PCR minor sampling.
• CQS rescue aka Torrent454 assembler implemented
• Can run for raw data, skipping consensus assembly step
• Multiple error models supported in variant caller
• Splitting long references to speed up and decrease memory footprint
• Spaced k-mers making initial reference mapping more robust
• Calling indels (no call quality is provided though)

MAGERI first release

1.0.0

Remove some obsolete stuff from pom

Mature pipeline

• Flexible de-multiplexing: remove primers, extract UMIs from adapters, positional UMI extraction, pre-processed data (UMI in header)
• CLI supports both manual (FASTQ pair) and project (set up via json) scale analysis
• Supports tabular, binary and VCF/SAM output
• Alignment is performed using hybrid k-mer + SmithWaterman algorithm, selected for its accuracy as MIG assembly greatly decreases the number of mappings required. MAPQ score is computed as k-mer overlap information measure
• Probabilistic error model for variant quality based on background error frequency

First release

Includes all basic features. Currently tested to support libraries generated using multiplex PCR.