Input preparation¶

Module 0 validates read files and detects paired-end or interleaved layouts. Supported extensions are .fq, .fastq, .fq.gz, and .fastq.gz.

Use one of these naming patterns:

reads_dir/
├── SampleA_R1.fastq.gz
├── SampleA_R2.fastq.gz
├── SampleB_1.fq
├── SampleB_2.fq
└── SampleC_interleaved.fastq

SAMWISE derives each sample ID from the complete filename prefix before the recognized read suffix. For the example above, the sample names are SampleA, SampleB, and SampleC. Illumina-style paired reads are supported, including sample_S1_L001_R1_001.fastq.gz and sample_S1_L001_R2_001.fastq.gz; their sample ID is sample_S1_L001.

This is prefix-based normalization: punctuation in the prefix is retained and the implementation does not truncate a sample ID at the first period. For example, these are distinct sample IDs:

Sample.1_interleaved.fastq
Sample.2_interleaved.fastq

They normalize to Sample.1 and Sample.2. Make sure each sample has a unique prefix before _R1, _R2, _1, _2, or _interleaved, and do not provide duplicate files for the same normalized sample and read role.

The root README contains an older statement that periods are removed from sample IDs. The current Module 0 implementation retains them, as described here.

For coassembly, provide a manifest that maps input samples to groups. An example is available at examples/coassembly_manifest_example.txt.