Scientific Methods — Targeted Comparison

Input & supplied targets

Supply 2–20 assemblies with distinct hashes and labels, and 1–50 gene symbols or locus identifiers. Target names are 1–64 characters, starting with an alphanumeric character; subsequent characters may also include underscore, dot, colon or hyphen. Names are deduplicated case-insensitively.

Uncompressed .fasta, .fa or .fna bacterial assemblies. Each file is limited to 25,000,000 bytes, 100,000–20,000,000 sequence bases and at most 10,000 contigs. Validation checks non-empty records, unique record identifiers and IUPAC DNA symbols.

This implementation accepts target names, not custom target FASTA sequences. It does not perform sequence-similarity target searches.

Annotation-based matching

Prokka annotates each assembly (--kingdom Bacteria, 2 CPUs). CDS, rRNA, tRNA and ncRNA GFF features are considered. A case-insensitive exact gene-symbol match is used first; if none is found in a genome, an exact locus-tag match is tried.

There is no alias, substring, fuzzy or product-description matching, and no identity/coverage threshold in this mode. PRESENT means an annotated feature matched the supplied name. NOT_DETECTED means no such annotation match was found; it does not prove biological absence.

Delivered outputs & interpretation

The report/ZIP include the target presence/absence table, genome-by-target matrix, feature evidence, per-genome hits, heatmap, assembly/annotation summaries, methods, tool versions, manifest and checksums. Calls depend on supplied target names and the annotation available for each assembly. Source assemblies are excluded.