LoginSign Up
← Back to all sparks
G

glydb

ANALYTICS
Velocity0.0

Glycan Structure Database

github.com/glycoverse/glydb ↗

glydb is pulling GlyGen's structure database onto the user's own disk.

glycomicsreference datadatabaseoffline accessr packages
◆Current state
glydb bundles glycan structure reference data and provides lookup helpers for compositions, structures, and species. The 0.6.0 release grew the bundled dataset to 19,436 GlyGen structures including non-intact glycans, and changed glytoucan_to_struc() to search that local data before falling back to the online GlyGen API. Version 0.4.0 added a confidence attribute that glyanno uses to rank candidate matches.
◆Where it's heading
The arc is from thin wrapper toward self-contained reference: the bundled data keeps growing, online lookups are demoted to fallbacks, and classification vocabularies are being adopted from GlyGen rather than invented locally. The other constant is chasing glyrepr, whose structure representation has changed enough times that regenerating the bundled indexes is a recurring release note.
◆Prediction
Expect the bundled dataset to track further GlyGen and GlyTouCan releases, with the online API path continuing to narrow to accessions the local data does not cover.

◆Recent moves

  1. 1mo ago

    19,436 bundled structures and offline-first lookup

    The bundled dataset grows to 19,436 GlyGen structures including non-intact glycans, and glytoucan_to_struc() checks it before calling the GlyGen API. Classification also moves to GlyGen's own vocabulary, covering HMO, GSL, GAG, GPI, and reducing-end O-linked subtypes.

    View source ↗
  2. 2mo ago

    glytoucan_to_struc() resolves accessions via GlyGen

    A new function fetches GlyTouCan accessions from the GlyGen API and parses the returned IUPAC strings into structures, plus a refresh to GlyTouCan v2.11.1. One release later this same function is rewired to prefer local data.

    View source ↗
  3. 4mo ago

    Documentation corrected to match data version

    A documentation fix stating the correct bundled data version. No code change.

    View source ↗
  4. 6mo ago

    Confidence scores and monosaccharide range filtering

    glydb_compositions() and glydb_structures() gain a mono_range parameter and return a confidence attribute, with a matching column in the bundled data. The attribute exists specifically so glyanno can rank candidate annotations rather than treat all matches as equal.

    View source ↗
  5. 6mo ago

    Bundled data updated to GlyGen v2.10.1

    A routine refresh of the internal dataset to a newer GlyGen release. Content changes, behavior does not.

    View source ↗
  6. 6mo ago

    Compatibility with glyrepr 0.10.0

    An adaptation to a glyrepr representation change. One of several such releases, which is why 0.6.0 had to regenerate the bundled structure indexes.

    View source ↗