R and Python · one verdict

biobouncer

A gate for biological inputs. It checks IDs at the door.

One small API answers "is this a valid identifier?" for gene symbols, ontology terms, variant formats, and database accessions across 50 sources. R and Python return the same verdict for the same input, enforced by a shared conformance corpus.

R docs → Python docs → Demo notebooks →
biobouncer hex logo: a bouncer with sunglasses and a checked ID badge

Two doors

One bouncer, two doors

Same ID, same answer. Pick the language you work in.

R

The R package

Function reference, a getting-started vignette, and recipes, built with pkgdown.

install.packages("biobouncer",
  repos = "https://samuelbharti.r-universe.dev")
Enter the R door →
 Python

The Python package

Guide, examples, a sources cookbook, and API reference, built with MkDocs.

pip install biobouncer
Enter the Python door →

Checking modes

Three ways to check

One argument, how, decides how strict and how online the check is.

patternIs it well-formed? Offline, deterministic.
cacheDoes it exist in a pinned snapshot? Offline, reproducible.
remoteDoes it exist right now in the source? Live.

How it works

A column in, a verdict out

A messy column goes in; one labeled verdict comes out, identical in R and Python.

biobouncer takes a messy column of biological identifiers, checks each through one gate with the checking modes, and returns one labeled verdict that is the same in R and Python.
Validate, normalize, and repair a whole column, not just one field.