Skip to content

Latest commit

 

History

History
656 lines (440 loc) · 22.1 KB

NEWS.md

File metadata and controls

656 lines (440 loc) · 22.1 KB
Raw

spatialLIBD 1.17.10

BUG FIXES

  • registration_wrapper() now automatically handles the scenario where k = 2 by not using registration_stats_anova() and providing an apporpriate warning. Implemented by @lahuuki at #86.

spatialLIBD 1.17.9

BUG FIXES

  • read10xVisiumWrapper() is now able to detect the GTF file used by SpaceRanger for version 3.0.0+. Implemented by @nick-eagles at #88.

spatialLIBD 1.17.6

BUG FIXES

  • Fixed the bug reported by @lahuuki about vis_grid_clus() not handling logical() cluster variables. See #80. To resolve this, sort_clusters() and get_colors() had to change internally. Examples and documentation for both functions have now been updated to showcase what happens when you provide a logical() vector as an input.

spatialLIBD 1.17.5

NEW FEATURES

spatialLIBD 1.17.3

NEW FEATURES

spatialLIBD 1.15.2

SIGNIFICANT USER-VISIBLE CHANGES

  • vis_gene() now has a multi_gene_method argument which provides 3 methods for combining multiple continuous variables: z_score, pca, and sparsity. These options can now be used with run_app() (the interactive websites). These methods are further illustrated and documented in a new vignette available at https://research.libd.org/spatialLIBD/articles/multi_gene_plots.html. This work was contributed by @Nick-Eagles.

spatialLIBD 1.13.6

NEW FEATURES

  • vis_clus_p(), vis_clus(), and vis_grid_clus() now all use implement the na_color argument that was present in the vis_gene() functions. This resolves #43 by @boyiguo1.

spatialLIBD 1.13.5

NEW FEATURES

  • run_app() now has a auto_crop_default argument set to TRUE by default. It can be turned off in cases where you are displaying images that do not follow the expected Visium grid dimensions, such as manually stitched images that you don't want to automatically crop.

spatialLIBD 1.13.4

NEW FEATURES

  • Added fetch_data("spatialDLPFC_Visium_example_subset") which is a subset of 3 samples with only the lowres images that can be used for example / tutorial purposes.

spatialLIBD 1.13.2

NEW FEATURES

  • Louise A. Huuki-Myers @lahuuki added a vignette explaining the spatial registration process and all related functions. See #46 for the full pull request.

spatialLIBD 1.11.13

SIGNIFICANT USER-VISIBLE CHANGES

  • The vignette now has a section describing the data from the spatialDLFPC, Visium_SPG_AD, and locus-c projects that were done by members of the Keri Martinowich, Kristen R. Maynard, and Leonardo Collado-Torres LIBD teams as well as our collaborators.

spatialLIBD 1.11.12

SIGNIFICANT USER-VISIBLE CHANGES

  • fetch_data("Visium_SPG_AD_Visium_wholegenome_spe""), fetch_data("Visium_SPG_AD_Visium_targeted_spe"), fetch_data("Visium_SPG_AD_Visium_wholegenome_pseudobulk_spe"), and fetch_data("Visium_SPG_AD_Visium_wholegenome_modeling_results") have been added. Use this to access data from the https://github.com/LieberInstitute/Visium_SPG_AD project.

spatialLIBD 1.11.11

SIGNIFICANT USER-VISIBLE CHANGES

spatialLIBD 1.11.10

BUG FIXES

spatialLIBD 1.11.9

SIGNIFICANT USER-VISIBLE CHANGES

  • gene_set_enrichment() now internally uses fisher.test(alternative = "greater") to test for odds ratios greater than 1. Otherwise odds ratios of 0 could be significant.

spatialLIBD 1.11.4

SIGNIFICANT USER-VISIBLE CHANGES

  • Several changes were made to the default plotting aspect of vis_gene(), vis_clus() and related plotting functions. This was done with input from @lahuuki and @nick-eagles and is described in more detail at https://github.com/LieberInstitute/spatialLIBD/commit/8fa8459d8fa881d254824d43e52193bf2c3021c0. Most noticeably, the aspect ratio is no longer stretched to fill the plotting area, the NA values will be shown with a light grey that has alpha blending, and the position of the legends has been made consistent between the plots.

spatialLIBD 1.11.3

NEW FEATURES

spatialLIBD 1.9.18

BUG FIXES

  • Fixed a bug related to edgeR::filterByExpr() inside of registration_pseudobulk().
  • Moved the min_ncells filtering step to registration_pseudobulk() rather than registration_wrapper() since you should drop low ncells before using edgeR::filterByExpr().

spatialLIBD 1.9.15

BUG FIXES

  • Fixed some bugs in registration_stats_anova() in cases where we only had two different unique values to compute F-statistics with, when we need at least
  • Made some parts of registration_stats_anova() and registration_stats_pairwise() more flexible.
  • registration_model() now provides a more informative error message when you have an empty factor level, thus leading to a non-full rank model matrix.

spatialLIBD 1.9.12

NEW FEATURES

  • Added functions for computing the modeling statistics used by the spatial registration process. See registration_wrapper() and related functions.
  • Added a function for using the output of layer_stat_cor() and for labeling the clusters. This can help interpret the spatial registration results. See annotate_registered_clusters() for more details.

spatialLIBD 1.9.11

BUG FIXES

  • Fixed bugs in gene_set_enrichment() for reverse = TRUE reported by @sparthib.
  • Added a reverse option on the shiny app under the gene set enrichment tab, that we tested with the example spe data.

spatialLIBD 1.9.10

SIGNIFICANT USER-VISIBLE CHANGES

  • Improved the automatic color palette selector when you switch discrete variables. It also now supports the ManualAnnotation option.
  • Discrete variable (cluster) legend is no longer duplicated under the clusters interactive tab.
  • You can now search the model test, which helps if you have lots of tests to choose from (this most likely occurs when you are looking at the pairwise results).

spatialLIBD 1.9.9

SIGNIFICANT USER-VISIBLE CHANGES

  • Made the shiny application more memory efficient in different areas.
  • Changed the default point_size from 1.25 to 2.
  • Added the option to show or hide the spatial images on the grid panels in the shiny web application. Turn off by default since it is more efficient.

spatialLIBD 1.9.5

BUG FIXES

  • Fix #41. Reported by @abspangler13. Now the gene selector changes automatically when you change the 'model results' (model type) or 'model test' inputs. The gene selector is now only shown inside the 'model boxplots' panel since it only affects that one.

spatialLIBD 1.9.4

BUG FIXES

  • Fix #40. Reported by @Erik-D-Nelson.

spatialLIBD 1.9.3

BUG FIXES

  • Added a more informative error message when 'stats' does not have ENSEMBL gene IDs as the rownames(). Reported by @abspangler13 and @sparthib at #33 (comment)

spatialLIBD 1.7.19

SIGNIFICANT USER-VISIBLE CHANGES

  • Documentation of the layer-level data panel at run_app() has been significantly increased. You can now also visualize more than 2 reduced dimensions computed on the pseudo-bulk level data (layer-level for the Maynard et al, Nature Neurosci, 2021 data).
  • Users can now control the font and point size on the reduced dimension plots, as well as the overall font size on the model boxplots.
  • Image edit scenarios you might be interested in for having a uniform color background image are now documented; for example if you want a white or black background, or actually any valid R color name or color HEX value.

spatialLIBD 1.7.18

SIGNIFICANT USER-VISIBLE CHANGES

  • run_app() now offers the option to chose any of the paletteer::paletteer_d color palettes for discrete variables.
  • Polychrome has been replaced as a dependency by paletteer. Note that Polychrome::palette36 is still the default.
  • run_app() now looks for columns that end with '_colors' in their name which can be used to pre-specify colors for any companion variables. For example if you have spe$my_groups and spe$my_groups_colors then the second one can specify the colors that will be used for visualizing spe$my_groups. This makes specifying default colors more flexible than before, and the user is still free to change them if necessary.

spatialLIBD 1.7.17

BUG FIXES

  • Fix bugs in layer_boxplot() where it was too specific to the Maynard et al 2021 data. We have made it more flexible now.
  • Made the y-axis space more dynamic in gene_set_enrichment_plot() and layer_matrix_plot().

spatialLIBD 1.7.16

BUG FIXES

  • Fixed a bug in sig_genes_extract() when there's only one set of t-statistics or F statistics to extract.

spatialLIBD 1.7.12

SIGNIFICANT USER-VISIBLE CHANGES

  • The visualization functions vis_*() of SpatialLIBD in this version match the Bioconductor 3.15 version of SpatialExperiment. Note that if you used SpatialExperiment::read10xVisium(), the names of the spatial coordinates changed at https://github.com/drighelli/SpatialExperiment/commit/6710fe8b0a7919191ecce989bb6831647385ef5f and thus you might need to switch them back if you created your SpatialExperiment object before this change. You can do so with spatialCoordsNames(spe) <- rev(spatialCoordsNames(spe)). read10xVisiumWrapper() uses SpatialExperiment::read10xVisium() internally, so this change on SpatialExperiment would then also affect you.

spatialLIBD 1.7.11

NEW FEATURES

  • Now layer_stat_cor() has the top_n argument which can be used for subsetting the marker genes prior to computing the correlation as part of the spatial registration process.

spatialLIBD 1.7.10

NEW FEATURES

  • Added the add_key() function to reduce code duplication and resolve #31.

spatialLIBD 1.7.9

NEW FEATURES

spatialLIBD 1.7.7

BUG FIXES

  • Fixed a bug where the using the left-mouse click was not working for annotating individual spots under the "gene (interactive)" tab.

spatialLIBD 1.7.6

NEW FEATURES

  • vis_gene_p(), vis_clus_p() and all related functions now have an argument point_size which lets you control how big the points are plotted. This can be useful for visualization purposes.
  • The shiny app now has an input controlling the point size. If you increase it to say 5, then if you zoom in the clusters (interactive) panel, you can see larger spots when zooming in.
  • These features are related to #28 although the spot diameter is still not the true spot diameter. However, now you have more flexibility for visualizing the spots.

spatialLIBD 1.7.5

NEW FEATURES

spatialLIBD 1.7.4

BUG FIXES

  • vis_gene() and vis_grid_gene() now support geneids that are found in the rownames(spe). This makese these functions more flexible.
  • vis_grid_gene() and vis_grid_clus() now have the sample_order argument which gives you more control in case you want to plot a subset of samples. This should also reduced the memory required as discovered at LieberInstitute/spatialDLPFC#45.

spatialLIBD 1.7.3

NEW FEATURES

  • Added support for more than one background picture per sample. This was done through the new argument image_id. Resolves #25.
  • Added options for side by side visualization of the background image and the clusters or gene expression values in the static versions. Resolves #19.
  • Allow changing the transparency level of the spots with the alpha argument. Resolves #20.
  • Add support for image manipulation with the magick package. Adds functions img_edit(), img_update() and img_update_all() as well as new features on the web application. Resolves #21.
  • Added support for more control over the gene color scale and in the web application also added support for reversing the order of the scale. Resolves #22 and #23.
  • Added export_cluster() and import_cluster() to help export/import clustering results instead of having to save large spe objects when exploring different clustering methods.
  • Added locate_images() and add_images() for adding non-standard images to a spe object.

spatialLIBD 1.7.2

BUG FIXES

  • Fixed an issue introduced by newer versions of shiny. This version of spatialLIBD works with shiny version 1.7.1, though it's likely backwards compatible. Resolves #24.
  • Fix an issue where as.data.frame(colData(spe)) uses check.names = TRUE by default and then changes the column names unintentionally.

spatialLIBD 1.7.1

NEW FEATURES

  • Added read10xVisiumWrapper() and related functions that make it easier to read in the SpaceRanger output files and launch a shiny web application using run_app(). These new functions read in the analysis output from SpaceRanger by 10x Genomics, in particular, the clustering and dimension reduction (projection) results.

spatialLIBD 1.3.19

SIGNIFICANT USER-VISIBLE CHANGES

  • spatialLIBD has been updated to work with SpatialExperiment version 1.1.701 which will be released as part of Bioconductor 3.13. This changes internal code of spatialLIBD which will work with any objects created with SpatialExperiment version 1.1.700.

spatialLIBD 1.3.16

SIGNIFICANT USER-VISIBLE CHANGES

spatialLIBD 1.3.15

SIGNIFICANT USER-VISIBLE CHANGES

  • We now use plotly::toWebGL() to make the web application more responsive.

spatialLIBD 1.3.14

SIGNIFICANT USER-VISIBLE CHANGES

  • The documentation and help messages shown in the web application have been revamped and improved.

spatialLIBD 1.3.12

NEW FEATURES

  • We added a new vignette that shows how you can use spatialLIBD with any 10x Genomics Visium dataset processed with spaceranger. The vignette uses the publicly available human lymph node example from the 10x Genomics website.

spatialLIBD 1.3.3

NEW FEATURES

  • Overall the package has been updated to use SpatialExperiment version 1.1.427 available on Bioconductor 3.13 (bioc-devel). Several functions were re-named such as sce_image_gene_p() now has a shorter name vis_gene_p(). This update also changes these visualization functions to ONLY support SpatialExperiment objects instead of the original modified SingleCellExperiment objects.
  • Updated citation information to reflect that https://doi.org/10.1038/s41593-020-00787-0 is now public. Also added a link on the README to https://doi.org/10.6084/m9.figshare.13623902.v1 for the manuscript high resolution images.

spatialLIBD 1.1.7

NEW FEATURES

  • The functions sce_image_gene_p(), sce_image_gene(), sce_image_grid(), sce_image_grid_gene(), sce_image_clus(), sce_image_clus_p(), geom_spatial() now work with VisiumExperiment objects thanks to the new function read_image() and ve_image_colData(). This work was done by Brenda Pardo and Leonardo.

spatialLIBD 1.1.5

NEW FEATURES

  • fetch_data() takes the data from sce object and creates a VisiumExperiment object containing these data thanks to the function sce_to_ve(). VisiumExperiment object can be obtained with fetch_data("ve"). This work was done by Brenda Pardo and Leonardo.

spatialLIBD 1.1.4

NEW FEATURES

  • fetch_data() now uses BiocFileCache() when downloading the data from Dropbox.

spatialLIBD 0.99.14

SIGNIFICANT USER-VISIBLE CHANGES

  • Added the function enough_ram() which is used to control the execution of examples. If it fails when using fetch_data("sce") then fetch_data() will show a warning.
  • fetch_data(type = "sce_example") is now supported and used visibly in the vignette, eliminating the need for eval = FALSE chunks. This should enable testing the vignette code on the Bioconductor Single Package Builder on Windows (max 2.5 GB of RAM available).

BUG FIXES

  • Fixed the example in get_colors().
  • Fixed layer_stat_cor_plot() for when min and/or max are specified.

spatialLIBD 0.99.13

SIGNIFICANT USER-VISIBLE CHANGES

spatialLIBD 0.99.12

BUG FIXES

spatialLIBD 0.99.11

BUG FIXES

spatialLIBD 0.99.10

SIGNIFICANT USER-VISIBLE CHANGES

BUG FIXES

spatialLIBD 0.99.9

SIGNIFICANT USER-VISIBLE CHANGES

spatialLIBD 0.99.8

SIGNIFICANT USER-VISIBLE CHANGES

spatialLIBD 0.99.7

BUG FIXES

spatialLIBD 0.99.6

SIGNIFICANT USER-VISIBLE CHANGES

  • Add mirrors for the shiny app and change the main location.

spatialLIBD 0.99.5

SIGNIFICANT USER-VISIBLE CHANGES

  • Make fetch_data() more flexible. Should now work when the data is absent.

spatialLIBD 0.99.4

BUG FIXES

  • Fix Travis badges
  • Fix Kristen's name on the vignette
  • Add the same welcome information to the top of the vignette, since this will be what Bioconductor users see first. Basically, we have made sure that users will see the same information first regardless if they find the package README, open the shiny app, or find the package vignette.

spatialLIBD 0.99.3

SIGNIFICANT USER-VISIBLE CHANGES

  • Further refine the READMEs (pkg and shiny). They now include the list of links to the raw 10x Genomics files as well as a short description of the project at the top. This was in response to feedback by Andrew Jaffe.

spatialLIBD 0.99.2

SIGNIFICANT USER-VISIBLE CHANGES

  • Update main package READMEs to reflect the changes to the shiny web app README.md.

spatialLIBD 0.99.1

NEW FEATURES

  • Added Kristen R Maynard to the DESCRIPTION file.
  • Improved the shiny app page footer.
  • Moved around the documentation and added a new main tab with an overview in response to the feedback by Stephanie Hicks.

spatialLIBD 0.99.0

NEW FEATURES

  • Added a NEWS.md file to track changes to the package.
  • First full version of the package to be submitted to Bioconductor. Note that the ExperimentHub::ExperimentHub() functionality won't work until they approve the package. However, for now fetch_data() has a backup mechanism in place.
  • Submitted to Bioconductor here.