To do

1. Dependency Review: Considering the removal of the GEOquery package, as it may no longer be necessary for preparing GEO series data.
2. Plot Customization: All plot* functions will be enhanced to allow greater user customization.
3. New S3 Class: Development of a new S3 class, similar to dgeList, which will offer improved customization and performance.

Version 1.3.0 Changes

• all exported functions follow the snake_case naming style; camelCase names are kept as deprecated aliases and will be removed in the next version (lintr config now enforces snake_case).
• list2dt() and list2graph() are the exception and keep their original names: the 2-for-“to” idiom matches base R (list2DF(), list2env()) and both names satisfy the snake_case linter, so renaming them would have broken a released name for no benefit. The briefly considered list_to_dt()/list_to_graph() never reached a release.
• function parameters renamed to snake_case as well (e.g., avg_log2fc_threshold, top_cell_n, tissue_class, tissue_type, min_count, min_unique_n, ignore_case, group_column, sample_info, feature_info, data_text, gsea_param, ticks_size, fgsea_res).
• prepare_tcga() returns expr_count and expr_fpkm instead of exprCount and exprFpkm, so every field of its two tables is snake_case. The old names still work and warn; they will stop working in 1.4.0.
• upgrade the built-in annotation database from CellMarker 2.0 to CellMarker 3.0 (418,139 entries; only the columns used by the package are kept, see data-raw/cellmarker.R).
• tuneParameters() now stores the annotation in the meta.data column “CellMarker3.0” instead of “CellMarker2.0”.
• matchCellMarker2() renamed to match_ref() because it also supports custom reference datasets. The old name is deprecated and will be removed in the next version.
• the built-in annotation dataset is renamed from cellMarker2 to cellMarker3 (internal; reflects the CellMarker 3.0 source).
• match_ref() ranks candidate cell types by uniqueN (breadth of matched markers) with N as tie-breaker instead of by N alone; a single heavily reported marker could otherwise outweigh a cell type matching on dozens of markers.
• match_ref() reports the detection rate (pct.1) of each matching marker in the new pct_with column, aligned with ordered_symbol and NA when the input has no pct.1. It is reported for auditing and does not affect the ranking; see the min_pct argument below if you want detection rate to influence the annotation.
• match_ref() gains a min_pct argument: markers whose detection rate in the cluster they were found for (pct.1) is below it are dropped before matching, so an annotation cannot rest on genes that are barely detected. It defaults to NULL (no filtering, i.e. the previous behaviour) and is ignored, with a message, if the input has no pct.1 column. Note that n is applied after this gate, so the markers used are not a subset of the unfiltered ones. Seurat::FindAllMarkers(min.pct = ) cannot replace it: that argument gates which genes are tested in either population, so a gene detected in a handful of cells can still end up as a positive marker.
• plot_possible_cell() gains a value argument: "N" (the default, unchanged), "uniqueN" (the measure match_ref() ranks candidates by) or "pct", which draws the share of matching markers that are actually detected at min_pct as tiles labelled with uniqueN — the evidence check for an annotation. value = "pct" errors when the input carries no detection rate. min_unique_n now includes candidates matching exactly that many markers; it previously required more than that.
• available_tissue_type() gains a tissue_class argument, so the tissue types recorded under one class can be listed (available_tissue_type("Human", tissue_class = "Blood")) before both are passed to match_ref(). A class the species does not have is an error rather than an empty result: tissue_class and tissue_type are two labels recorded per database entry, not a hierarchy, so whether a pair exists is worth checking rather than assuming.
• match_ref() warns when the tissue filters leave an empty reference set, instead of silently returning a table with no candidate in it. A class and a type that never occur together on a database entry select nothing, which is easy to write by accident because the two fields look like a hierarchy; available_tissue_type() above is how to check.
• the single-cell vignette and the example script it mirrors bound the annotation search to tissue_class = c("Blood", "Bone marrow"), with the reasoning: the default searches every organ, and bounding the class is what makes the candidate list readable, while bounding tissue_type as well removes evidence rather than correcting for anything.
• the deprecated aliases report “deprecated as of 1.3.0” when called and are scheduled for removal in 1.4.0.
• the startup message now names the two changes above — the CellMarker 3.0 database, which changed underneath the same function names and arguments, and the snake_case names — instead of announcing unnamed “significant breaking changes”, and it is printed only in an interactive session, so R CMD check, CI and Rscript stay quiet.

Version 1.2.3 Changes

• dont run Seurat::DotPlot in vignette to avoid upstream upgrade’s affects

Version 1.2.2 Changes

• fix vignette name in package startup message

Version 1.2.1 Changes

• make each function do a simple task instead of integrating all to a function(check_marker, plotSeuratDot).
• support using custom dataset to annotate and allow user to set the threshold(matchCellMarker2).
• use more convenient input in finsert.
• guess user’s typo input (get_marker).
• update docs and hints.
• set the minimal data.table version to 1.15.0
• support generating plots for each cell type in plotSeuratDot.

Version 1.1.1 Changes

• support argument tissueType and tissueClass in single cell related function.
• fix some typo errors in example-single-cell.R
• update some miscellaneous functions.

Version 1.1.0 Changes

Enhancements

• Artist Class: Now includes a default data argument for data exploration. All commands and results are saved, allowing users to revisit previous analyses.
• Bulk RNA-seq Vignettes: Updated for better usability.
• get_marker() Updates: Added extra messages when searching for undefined cells in the database, improving clarity for users.

Bug Fixes

• prepare_geo() Fix: Non-character ID columns in GPL data are now converted to character to prevent misinterpretation of numeric IDs. Expression data can now be read directly from supplementary files without needing to download them locally.
• plotORA() Fix: Fixed an issue where the plot would fail to display a legend when mapping a variable to the fill aesthetic.

Version 1.0.1 Changes

• NEWS.md File: Added to document changes and updates.
• macOS Compatibility: Fixed an error in uniprot_id_map() on the latest macOS during example usage.
• Vignette Optimization: Vignettes have been updated and optimized for improved performance and usability.