fillpattern
Pattern fills for ggplot2, hardened against the ways users write sizes
Comparison · Analytics
tglkmeans vs UCell
A side-by-side editorial comparison of tglkmeans and UCell — release velocity, themes, recent moves, and the top alternatives to consider.
Shared themes:r-packagesingle-cell
tglkmeans vs UCell: at a glance
What is tglkmeans?
A k-means implementation that just told users their Spearman clustering on missing data was wrong
tglkmeans is a multi-core k-means implementation with seeding, aimed at single-cell and other large matrix workloads. Version 0.4.0 flipped the id_column default and moved to R's random number generator, 0.5.x added count-matrix downsampling and fixed id handling, and 0.6.3 in May 2026 is a correctness release: Spearman distance was ranking missing values as the largest value instead of dropping them, and predict_tgl_kmeans() with Euclidean distance did not reproduce the training metric when a cluster center had a missing dimension.
What is UCell?
A rank-based gene signature scorer that has grown by adapting to whatever object format single-cell R uses next
UCell scores gene signatures in single-cell data using a rank-based metric that is robust to dataset composition. Its release history reads as a sequence of ecosystem accommodations: Bioconductor submission in 2.0, SmoothKNN() for k-nearest-neighbor smoothing of scores in 2.2, smoothing applied directly to expression slots in 2.4, Seurat v5 assay compatibility in 2.6, multi-layer Seurat v5 objects in 2.8, and a missing_genes parameter in 2.14 that lets callers impute or skip signature genes absent from the data. Version 2.16 tracks Bioconductor 3.23 and points at a new publication and a Python implementation, pyUCell.
tglkmeans vs UCell: editorial side-by-side
tglkmeans
0.0ANALYTICS
A k-means implementation that just told users their Spearman clustering on missing data was wrong
◆ Current state
tglkmeans is a multi-core k-means implementation with seeding, aimed at single-cell and other large matrix workloads. Version 0.4.0 flipped the id_column default and moved to R's random number generator, 0.5.x added count-matrix downsampling and fixed id handling, and 0.6.3 in May 2026 is a correctness release: Spearman distance was ranking missing values as the largest value instead of dropping them, and predict_tgl_kmeans() with Euclidean distance did not reproduce the training metric when a cluster center had a missing dimension.
◆ Where it's heading
The package handles missing data across three distance metrics, and 0.6.3 shows those paths had drifted apart — Spearman behaved unlike Euclidean and Pearson, and prediction behaved unlike training. Both fixes change results on affected data, and the release notes are careful to bound exactly where: Spearman on data with NAs changes, complete data does not. Performance work runs alongside, with the dense per-thread vote matrix removed from the reassignment step.
◆ Prediction
With the metric paths now aligned on missing-value handling, further work is more likely to target the parallel reassignment internals than the distance semantics.
UCell
0.0ANALYTICS
A rank-based gene signature scorer that has grown by adapting to whatever object format single-cell R uses next
◆ Current state
UCell scores gene signatures in single-cell data using a rank-based metric that is robust to dataset composition. Its release history reads as a sequence of ecosystem accommodations: Bioconductor submission in 2.0, SmoothKNN() for k-nearest-neighbor smoothing of scores in 2.2, smoothing applied directly to expression slots in 2.4, Seurat v5 assay compatibility in 2.6, multi-layer Seurat v5 objects in 2.8, and a missing_genes parameter in 2.14 that lets callers impute or skip signature genes absent from the data. Version 2.16 tracks Bioconductor 3.23 and points at a new publication and a Python implementation, pyUCell.
◆ Where it's heading
Two threads run through this. The scoring algorithm itself has barely changed — the rank-based core is stable, and 2.14's reformatting to gene indices rather than string matching is a speed change, not a method change. What does change constantly is object-format compatibility, which is the tax of living between Seurat and SingleCellExperiment. The pyUCell reference in 2.16 is the first sign of the method reaching beyond R, though these notes say nothing about its scope.
◆ Prediction
The cadence is locked to Bioconductor's twice-yearly release train, so the next version will most likely accompany Bioconductor 3.24 with whatever Seurat or SingleCellExperiment changes it brings.
Alternatives to tglkmeans and UCell
Other Analytics products tracked by Sparkpulse, ranked by recent ship velocity. Each card links to a full editorial trajectory and lets you pivot into a head-to-head comparison with either tglkmeans or UCell.
gcube
gcube's recent releases are all packaging metadata, not simulation code
cubist
The R port of Quinlan's Cubist gets reproducibility fixes, not new modelling
ggstats
ggstats keeps widening what a coefficient or Likert plot can be
ecodive
ecodive rebuilt itself into a broad diversity-metric library, breaking as it went
simStateSpace
State-space data simulation for R, filled in one function at a time
See all tglkmeans alternatives → · See all UCell alternatives →
Recent activity from tglkmeans and UCell
Latest ship moves from both products, interleaved chronologically. ⚡ = editorial spark.
- 2mo agotglkmeansSpearman metric no longer ranks missing values as the largest
- 3mo agoUCellTracks Bioconductor 3.23 and points to a Python port
- 9mo agoUCellUCell version 2.14
- 2y agoUCellUCell version 2.8
- 2y agotglkmeansFixes corrupted cluster ids and dropped dimnames
- 2y agotglkmeansAdds downsample_matrix() for count matrices
- 2y agotglkmeansBreaking: id_column defaults to FALSE, switches to R's RNG
- 2y agoUCellUCell version 2.6
- 3y agoUCellUCell version 2.4
- 3y agoUCellUCell version 2.2
Frequently asked questions
What is the difference between tglkmeans and UCell?
Both compete on the same themes — r-package, single-cell — within Analytics. tglkmeans and UCell are shipping at a similar cadence (velocity 0.0 vs 0.0, both within Sparkpulse's "active" band). See the at-a-glance table above for a side-by-side breakdown of velocity, recent sparks, and editorial themes.
Is tglkmeans better than UCell?
Sparkpulse doesn't pick a winner — we score release velocity, not feature parity. tglkmeans and UCell are shipping at a similar cadence (velocity 0.0 vs 0.0, both within Sparkpulse's "active" band). For your specific use case, the alternatives sections above list other Analytics products to evaluate alongside.
What are the best alternatives to tglkmeans?
Top tglkmeans alternatives in Analytics are ranked by recent ship velocity. Browse the "tglkmeans alternatives" section above for the current picks, or visit /alternatives/tglkmeans for the full list with editorial commentary on each.
What are the best alternatives to UCell?
Top UCell alternatives in Analytics are ranked by recent ship velocity. Browse the "UCell alternatives" section above for the current picks, or visit /alternatives/ucell for the full list with editorial commentary on each.