Skip to the options

Reference

Every option, output file and cell state on one page.

Options Output files Cell states Rescue types Go deeper

Options

Only -p, -r and -g are required. recore --help prints the same list.

Option	Default	What it does
Inputs
`-p`, `--presence-absence`	required	The presence/absence matrix: Panaroo, Roary, PPanGGOLiN or PanTA CSV, or PIRATE's gene family table.
`-r`, `--pan-reference`	required	One representative sequence per cluster, named like the matrix rows.
`-g`, `--genomes`	required	Folder of assemblies, one per matrix column, named after it.
`-a`, `--annotations`	none	Folder of GFF files, named like the assemblies. Lets Recore tell split, paralog and start-variant hits apart.
`--tree`	none	Rooted Newick tree of the genomes, for gain and loss calls and the report's phylogeny.
`--reference-genome`	none	A matrix column (or its AS identifier) to audit as the reference.
`--mobile-element-reference`	none	FASTA of IS elements or transposases, for sequence support beside each rescue.
Search
`--min-identity`	90	Minimum % identity for a hit.
`--min-coverage`	80	Minimum % of the query a hit, or its merged fragments, must cover.
`--preset`	asm5	minimap2 preset: `asm5` up to 5% divergence and faster, `asm20` up to 20%.
`--min-prevalence`	0.1	Only search clusters present in at least this fraction of genomes. `0` searches all.
`--multi-allele`	off	Search every allele of a cluster, not only the longest. Slower, finds diverged alleles.
`--search-cache`	none	Keep search results in a folder; a rerun aligns only what changed.
Decisions
`--synteny`	off	Check each candidate's neighbouring genes before accepting it. Needs `-a`.
`--matrix-policy`	balanced	Which evidence reaches the corrected matrix: `strict`, `balanced` or `sensitive`.
Outputs
`-o`, `--output`	recore_output	Folder for every result.
`--write-html-report`	off	Write `recore_report.html`, the run in one offline page.
`--write-corrected-gff`	off	Write each GFF with the rescued genes added.
`--write-rescued-fasta`	off	Write the sequence of every gene put back, as nucleotides and protein.
Mobile elements
`--mobile-element-min-identity`	85	Minimum % identity for a mobile-element hit.
`--mobile-element-min-coverage`	50	Minimum % coverage for a mobile-element hit.
`--mobile-element-window`	1000	How far (bp) from a rescue a mobile element still counts.
Redundant clusters
`--nscluster`	off	Merge clusters that differ as proteins but are the same DNA.
`--nsc-engine`	minhash	`minhash` is built in; `mmseqs2` aligns, and needs MMseqs2 installed.
`--nsc-kmer-size`	21	k-mer size for the MinHash sketches.
`--nsc-threshold`	0.8	Containment needed to merge, with MinHash. Higher merges less.
`--nsc-min-seq-id`	0.9	Identity needed to merge, with MMseqs2.
`--nsc-coverage`	0.8	Coverage needed to merge, with MMseqs2.
Run
`-t`, `--threads`	4	Threads.
`-v`, `--verbose`	off	Debug logging: name matches, GFF parsing, synteny decisions.
`-h`, `--help`		Print this list.
`-V`, `--version`		Print the version.

No option matches.

Output files

Everything goes into the output folder. Your input files are never changed.

File	Written	What it is for
Start here
`gene_presence_absence_recore.csv`	always	The corrected matrix, in the input's layout, for downstream tools.
`gene_presence_absence_recore.Rtab`	always	The same matrix as 1/0, for pyseer and R.
`gene_presence_absence_recore_states.csv`	always	A state in every cell (see Cell states), for figures and audit.
`orthology_evidence.tsv`	always	One row per decision: evidence, state, matrix action and wording.
`recore_report.html`	`--write-html-report`	The whole run in one page that works offline.
`summary.txt`	always	Counts in plain text: gaps, rescues, core genome before and after.
Rescues and rejections
`rescue_report.tsv`	always	One row per rescued gene: type, identity, coverage, coordinates.
`synteny_rejected_report.tsv`	`--synteny`, when any	Candidates whose neighbours did not match.
`paralog_rejected_report.tsv`	when any	Paralog hits without synteny support.
`rescued_genes.fna`, `.faa`	`--write-rescued-fasta`	Sequences of every gene put back.
`corrected_gff/*.recore.gff3`	`--write-corrected-gff`	Each GFF with the rescued genes added.
Genes, clusters and genomes
`evolutionary_events.tsv`	always	What best explains each cluster: gain, loss, IS insertion, pseudogene, paralog.
`cluster_relationships.tsv`	always	Clusters that look like one gene split in two.
`cluster_components.tsv`, `cluster_component_members.tsv`	always	Groups of related clusters.
`cluster_locus_context.tsv`, `cluster_locus_summary.tsv`	with `-a`	Each cluster's neighbours, and whether it sits at one locus.
`cluster_synteny_blocks.tsv`	with `-a`	Five-gene neighbourhoods on each side, grouped per cluster.
`genome_qc.tsv`	always	Per genome: how many rescues and missing annotations.
`reference_genome_audit.tsv`	`--reference-genome`	Whether each cluster is really absent from the reference.
`reference_cluster_events.tsv`	`--reference-genome`, when any	Reference-absent clusters that look like a reference allele or fragment.
`nscluster_report.tsv`	`--nscluster`, when any	Clusters merged as the same DNA.
`run_manifest.json`	always	Version, inputs and parameters, for reproducibility.

Cell states

What gene_presence_absence_recore_states.csv says about each cell, and how the report draws it.

State	Meaning	Put back with `balanced`
`PRESENT_INPUT`	Present in your matrix.	already there
`STRICT_PRESENT`	Same locus, by annotation or by its neighbours.	yes
`ALIAS_CLUSTER`	The same gene, filed under another cluster.	yes
`SEQUENCE_ONLY`	The full-length gene is in the assembly.	yes
`MISSING_CDS_UNVALIDATED`	Full-length sequence where nothing is annotated, and no synteny to back it.	no
`PARTIAL_HOMOLOGY`	Only part of the gene.	no
`PSEUDOGENE`	The gene is there, but disrupted.	no
`HOMOLOG_PARALOG`	A related gene, not the same one.	no
`REJECTED_CONTEXT`	Its neighbours or the paralog check say otherwise.	no
`ABSENT`	Searched; nothing found.	no
`NOT_SEARCHED_LOW_PREVALENCE`	Not searched: the cluster is below `--min-prevalence`.	no
`NOT_SEARCHED_NO_CLUSTER_SEQUENCE`	Not searched: no sequence for the cluster.	no
`NOT_SEARCHED_NO_FASTA`	Not searched: no assembly for the genome.	no

input put back merged as the same DNA (--nscluster) kept out, with the reason absent not searched

Rescue types

What Recore saw at the place a missing gene was found, in rescue_report.tsv.

Type	What it means
`DROPPED`	The gene is in the assembly and no other cluster claims that place: the matrix lost it.
`MISSING_CDS`	The gene is in the assembly, but the GFF has no CDS there.
`SPLIT`	That place is another cluster's gene, at 95% identity or more: one gene split in two clusters.
`FRAGMENTED`	The gene is in pieces: across contig ends, an IS insertion or a large indel.
`ALIGNMENT_SPLIT`	Several alignments make up the gene, with no real break between them.
`START_VARIANT`	The same gene, starting at a different position.
`PARALOG`	A related gene from the same family, under 95% identity: not the missing one.
`PSEUDOGENE`	The gene is disrupted: an internal stop, a frameshift or a partial deletion.

Go deeper

The long versions, with every column and every rule, in the documentation.

TutorialThe example run, number by number. InputsWhat each tool writes, and how names match. Output filesEvery table, column by column. Matrix policiesThe three policies and the quality guard. The reportEvery pane and control of the viewer. ParametersEvery option, with what it changes. How it worksEach step of the pipeline. Evolutionary eventsHow an event is called, and how to word it. Limits and speedWhat Recore does not claim; run time and memory.