Skip to content

README.md

Latest commit

 

History

History
82 lines (75 loc) · 4.36 KB

README.md

File metadata and controls

82 lines (75 loc) · 4.36 KB
Raw

example_data — worked examples + required input structure per mode

phageFACTor runs in three input modes. The input files and folder layout differ per mode — this is the most important thing for a new user, so each is spelled out below with a small worked example. Auto‑detection (detect_input_type.py) reads input/fasta/ and picks genome, protein, or pangenome (bulk nucleotide CDS).

example_data/
├── genome_2campy/     # Example 1: genome mode  (2 Campylobacter prophages) — shipped, ready to run
├── protein_2citro/    # Example 2: protein / pre‑annotated‑CDS mode (2 Citrobacter prophages: phiNP, phiSM)
├── run_smoke_test.sh  # end‑to‑end check on Example 1 (WebAPI mode, no local DB)
└── pangenome/         # Example 3: pangenome mode (bulk nucleotide CDS, no genome coords) — worked example to add

Example 1 — GENOME mode (the simplest; 2 Campylobacter prophages)

Input = just the nucleotide genome FASTAs. Pharokka (Phanotate) calls the CDS, phold

  • FoldSeek annotate. Minimal input:
input/
├── fasta/
│   ├── CJIE1.fasta            # one nucleotide FASTA per prophage (>name on the header)
│   └── CJIE2.fasta
└── prophage_list.txt          # the prophage basenames, one per line (CJIE1 / CJIE2)

.input_type is written automatically (= genome). Run: bash submit_all.sh.

Example 2 — PROTEIN / pre‑annotated‑CDS mode (2 Citrobacter prophages: phiNP, phiSM)

Use when the CDS are predefined (e.g. host annotation quantified by RNAseq) — no pharokka. You provide the proteins + their coordinates/annotations. (Input files are not committed yet — see protein_2citro/README.md.) Required input:

input/
├── .input_type                       # contains the word: protein
├── fasta/
│   ├── <Prophage>_proteins.faa       # AA FASTA per prophage; header = gene id (locus_tag)
│   └── prophage_windows.csv          # prophage,genome_start,genome_stop,contig
├── gene_metadata_rich.csv            # the join table (schema below)
└── genome/
    └── source_genome.fasta           # host genome (multi‑contig OK) — for the GBK

gene_metadata_rich.csv columns (one row per CDS): prophage, contig, gene_id_gff, gene_id_fa, start, end, strand, aa_length, gff_product, emapper_desc, COG_category, preferred_name, is_hypo_gff, is_hypo_emapper, is_hypothetical

  • gene_id_fa must equal the protein FASTA header id (join key).
  • start/end are absolute host‑genome coords (windows can span the origin); contig is optional for a single-chromosome host (as in the citro example).
  • is_hypothetical (= hypo in the host annotation AND in eggNOG) defines the FoldSeek target set.
  • You build this table once from your host annotation (GFF + proteins.fa + eggNOG + the prophage coordinate windows); protein_2citro/ is a worked example of the result. Run: bash submit_all.sh (auto‑detects protein).

Example 3 — PANGENOME mode

Bulk nucleotide CDS with no per‑gene genome coordinates (e.g. a clustered pan‑proteome / Roary representative FASTA). Minimal input:

input/
└── fasta/
    └── <name>.fa          # one nucleotide CDS per record (Roary group in the header)

Leave INPUT_MODE unset (auto‑detect returns pangenome) or set INPUT_MODE=pangenome, then bash submit_all.sh. It translates the CDS to protein (clean pang_NNNNN ids + input/pangenome_id_map.csv), runs protein mode (minus prophage_windows.csv + genome/ → no GBK / no phynteny), and after curation builds 04_output/pangenome_reference.{csv,xlsx,faa} (one row per group + a DIAMOND/MMseqs‑ready FASTA) for homology‑based annotation transfer. A shipped worked example is to be added.

Expected output (any mode)

04_output/final_annotations_table.csv (one row per CDS) + 04_output/curation/review_suggested.csv (the few genes flagged for a human look) + 05_phynteny/final_annotations_integrated.csv after the phynteny step. run_smoke_test.sh checks the table's schema and annotated fraction; it runs WebAPI mode (no local DB) so a check needs no big database.

Status: both genome_2campy/ and protein_2citro/ (phiNP + phiSM, 173 proteins) are shipped and runnable. After a run you may commit a small final_annotations_table.csv subset under an expected/ folder as an illustrative reference.