Repository navigation
feat: NACHO 3.0.0 - #139
Merged
Merged
Conversation
# Conflicts: # tests/testthat/test-autoplot.R
# Conflicts: # DESCRIPTION # NEWS.md
PCL needs three usable positive controls, since a correlation on two points is always 1 and passed QC without assessing linearity. A rebuild now warns only about samples whose positive factor was usable before, so normalise() does not repeat what load_rcc() already said.
The other fixtures come from nSolver 4.0.0.3 and 2.3. GSE346526 holds RCC files from the most recent nSolver version on GEO, so the tests also cover data written today.
Whole-blood RNA raises LIG_NEG_A together with red blood cell miRNAs such as miR-451a, so most samples of a whole-blood study fail Ligation_NEG. GSE303669 and GSE317496, two independent PAXgene studies on GEO, show it. load_rcc() now warns when more than half of the samples fail, gives the miR-451a level, and points to relaxing the threshold rather than dropping the samples. The rule itself is unchanged.
The upload name went into file paths as given, and zip entries with ../ or absolute paths were extracted as given, which R before 4.5.1 does not sanitise. Upload names are now reduced to their base name, and an archive with an entry outside its folder is refused before extraction.
An id column named like a computed column, such as ID or Date, was dropped while building the object, and the validator then failed with an unrelated message. load_rcc() and as_nacho() now stop early and ask to rename the column.
A housekeeping gene name that matched no probe, such as a typo, gave a NaN House_factor and NaN normalised counts while QC still passed. load_rcc() and normalise() now stop and name the unknown genes.
dirname() gives forward slashes on Windows while the temporary folder uses backslashes, so the upload-name test failed there.
Rewrite the development section for readers of the package: one sentence per line, active voice, plainer wording, and the PlexSet fix listed under R/app.R now that inst/app/ only wraps it.
Drop fixes and speed-ups to features that are new in 3.0.0, and fold the behaviour that still matters into the matching feature entries.
The selected point kept growing into a dark disc that hid the box and the points next to it. It now keeps its fill, grows a little and gets a rose ring, which stays apart from the group and flag colours.
All app figures, the README animation and the quick-start plot now show the current app. The script clears the sample selection and the focus before the shots that should not show them.
An id column called status, reason, n_flags, lane, lane_status or <metric>_status was overwritten by nacho_qc(), so the sample ids read pass or fail. load_rcc() and as_nacho() now refuse those names.
GSE270837 was measured on an nCounter SPRINT, so the vignette now uses its binding density limits.
With Preset and Imaging both open, the sidebar ran far below the window. Each page now opens the threshold section that matches it, such as Imaging on QC metrics, and folds the others.
Take the screenshots 1192 pixels tall, so the sidebar and most pages fit, and give the animation a matching height instead of the gifski default that squashed its frames.
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
This brings NACHO 3 to
main, ready for the 3.0.0 release.NACHO 3 is a breaking release.
The object is now an S7 class, read with
nacho_counts(),nacho_samples(),nacho_probes()andnacho_qc().Quality control follows the nSolver thresholds by default, and a legacy preset gives back the NACHO 2 calls.
The report is built with Quarto, as HTML or a Typst PDF.
The app is rebuilt on bslib, with linked interactive plots, a guided tour, a background import with an upload preview, a sortable Samples table, and an Export page that also writes an R script of the analysis.
NEWS.mdlists the changes since 2.0.8, and the article "Moving from NACHO 2" maps the old code to the new.The branch holds 42 pull requests, each reviewed before it merged.
ggiraphandgdtoolsinSuggeststo the CRAN releases that fixes [Bug]girafe()rebuilds the Liberation font set even whenfont_setis supplied davidgohel/ggiraph#360 and fixes [Bug]font_family_exists()lists every system font on each call davidgohel/gdtools#84, so interactive plots build in 0.40.2 to 0.80.6 s instead of about 2 s.