LoginSign Up
← Back to all sparks
G

glyexp

ANALYTICS
Velocity0.0

Glycoproteomics and Glycomics Experiments

github.com/glycoverse/glyexp ↗

glyexp is retiring its own data container and handing the job to Bioconductor.

glycomicsbioconductordata containersbreaking changesr packagesmigration
◆Current state
glyexp is the container layer under the glycoverse stack, and it just changed what that container is. Versions 0.15.0 and 0.16.0 introduced GlycomicSE and GlycoproteomicSE as SummarizedExperiment subclasses, taught the dplyr-style verbs to operate on them, and then deprecated the legacy experiment() constructor and its accessors. Earlier releases in the window were narrower: as_pseudo_glycome(), a magrittr-free rewrite, and an offline standardize_variable().
◆Where it's heading
The package is moving from a bespoke object model to the Bioconductor one, and doing it in explicitly numbered stages tracked in a single issue (glyexp#15). Stage I added the subclasses as experimental; Stage II deprecated the old container and pushed the migration through ten sibling packages within days. The tidy manipulation verbs are being kept as the compatibility bridge, which suggests the dplyr surface is what the maintainer considers glyexp's actual contribution once the container is someone else's.
◆Prediction
Expect a Stage III release that removes the deprecated experiment() constructor and accessors outright, leaving GlycomicSE and GlycoproteomicSE as the only supported containers.

◆Recent moves

  1. 1mo ago

    glyexp deprecates its own container for SummarizedExperiment

    ⚡ SPARK

    This is the release where the migration stops being optional: the legacy experiment() constructor, conversion helpers, accessors, metadata helpers, and S3 methods are all deprecated, and the bundled example datasets are rebuilt on GlycomicSE and GlycoproteomicSE. The dplyr-style verbs now accept SummarizedExperiment objects directly, exposing dimensions as virtual .sample and .variable columns.

    View source ↗
  2. 1mo ago

    Experimental GlycomicSE and GlycoproteomicSE containers land

    Stage I of the container migration: the SummarizedExperiment subclasses arrive as experimental, with coercion and predicate helpers, and summarize_experiment() and as_pseudo_glycome() learn to handle them. glycan_type validation also widens to O, HMO, GSL, GAG, and GPI.

    View source ↗
  3. 1mo ago

    from_se() metadata fixes and dataset refresh

    Housekeeping ahead of the container work: from_se() preserves all SummarizedExperiment metadata, accepts traitomics and traitproteomics types, and errors when exp_type is missing everywhere. The bundled datasets are refreshed to the current glyrepr representation.

    View source ↗
  4. 4mo ago

    filter_obs() and filter_var() drop unused factor levels

    A single fix aligning filter_obs() and filter_var() with their documented behavior. Notable only because the same factor-level question resurfaces in 0.16.0, where the default flips and .drop_levels becomes opt-in.

    View source ↗
  5. 4mo ago

    as_pseudo_glycome() converts glycoproteomics to glycomics

    as_pseudo_glycome() lets glycoproteomics data be analyzed with glycomics methods, and the package drops magrittr for the native pipe. The conversion helper is the kind of cross-type plumbing that the later SummarizedExperiment subclasses generalize.

    View source ↗
  6. 5mo ago

    standardize_variable() drops its UniProt network dependency

    standardize_variable() is simplified to work offline, losing the fasta and taxid parameters and the UniProt.ws dependency in favor of a protein_site column. Removing a network round-trip from a core normalization step is the sort of dependency pruning that recurs across this stack.

    View source ↗