Pathogen Annotation Sheet¶

Every organism TaxTriage can flag, with the annotations that drive Microbial Categories and the TASS confidence score.

This table reads assets/pathogen_sheet.csv straight from the main branch of the pipeline repository when the page loads - it is not bundled into this site. Merge a change to that CSV and it shows up here on the next refresh, with no redeploy. Columns the sheet does not currently carry are left blank.

Using this table

Search matches organism names, synonyms, tax IDs and full lineage. Facets combine with AND across groups and OR within a group, and each group's counts reflect the other active filters. Click any row for the full record including references. Export CSV downloads exactly what you have filtered to - handy for building a custom --pathogens sheet.

Loading the pathogen sheet…

Column reference¶

Column	Meaning
`name`	Organism name, matched against the taxonomic classification output
`taxid`	NCBI Taxonomy identifier
`general_classification`	`primary`, `opportunistic`, `potential`, or `commensal` - see Microbial Categories
`status`	`established` where the pathogen - site association is well documented, `putative` where it is suggested but not confirmed
`high_consequence`	Flags organisms warranting immediate escalation regardless of abundance
`pathogenic_sites`	Body sites where detection is considered clinically meaningful
`commensal_sites`	Body sites where the organism is expected flora, which down-weights its score
`alternative_names`	Synonyms and former names, also searched during matching
`pathology`	Free-text disease association
`host_organism`	Host the annotation applies to
`kingdom` … `genus`	Lineage, used for genus-level rollups
`mol_type`	`dna` or `rna`, which determines the alignment path
`reference`	Literature supporting the annotation
`assembly_accession`	Default reference assembly downloaded for alignment

Requesting changes¶

Request changes above covers both directions: adding organisms the sheet does not carry, and correcting ones it does.

Adding organisms¶

Fill the form and choose Add another to queue as many as you like. Searching for something that is missing also offers to request it directly from the empty result, with the query prefilled.

Updating an existing entry¶

Switch to Update existing entries, pick the organism, and the form loads its current values. Change only what is wrong - the request records the fields you actually touched, as current → proposed, and says explicitly that everything else stays as it is. A short reason is required, since that is what a reviewer weighs. Opening any row and choosing Request an update to this entry starts the same flow with that organism already loaded.

Additions and updates can be queued together and go out as one issue.

Reviewing before you submit¶

Neither Review & open issue nor Review & download submits anything straight away. Both first show a confirmation window listing every staged entry: new organisms with their full field set, and updates as a table of just the changed fields, current beside proposed. Only Confirm hands the request over

to a prefilled GitHub issue, or to a Markdown file you can send privately if the request is sensitive. Back to editing returns you to the form with everything intact.

The request_type column follows the route: git-tracked for an issue, external-local for a download.

The sheet is a plain CSV, so changes can equally be pull requests against the pipeline repo. See Contributing for the required fields and review expectations.

To run with your own sheet instead of the bundled one:

nextflow run jhuapl-bio/taxtriage \
    --input samplesheet.csv \
    --pathogens /path/to/my_pathogen_sheet.csv \
    -profile docker