forked from HelenaLC/CATALYST
-
Notifications
You must be signed in to change notification settings - Fork 0
/
NEWS
148 lines (109 loc) · 6.36 KB
/
NEWS
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
changes in version 1.14.1 (2021-28-04)
preprocessing
+ bug fix in computeSpillmat() when 'interactions = "all"'
+ added unit tests for all combinations of 'interactions' and 'method'
changes in version 1.12.2
preprocessing
+ added argument 'FACS' to prepData(); dropping of non-mass channels
can now be ommitted explicitly via setting FACS = TRUE
+ allow coloring by clustering in plotScatter()
differential
+ added new visualization clrDR() to plot low-dimensional embedding
of centered log-ratios (CLR) on cluster-compositions across samples
+ added argument 'assay' in plotExprs(), plotNRS()
to allow specification of which assay data to use
+ depracated plotMedExprs() in favor of plotPbExprs() to allow flexibilty
in input data and summary statistic, aggregating by cluster-sample
(instead of by sample only), sizing by sample(-cluster) cell counts,
and to specify whether to include (jittered) points, boxplots or both
+ but fix in plotAbundances(); plotting failed
when colData()$sample_id was not of type factor or character
+ changed behavior of 'group_by' in plotAbundances();
not facetting the plot was previously impossible
+ added arguments 'col_clust,distance,linkage' in plotAbundances()
to allow sorting of samples according to hierarchical clustering
on cluster compositions when by = "sample_id"
changes in version 1.12.1
preprocessing
+ bug fix in prepData(): by default, non-mass channels (time, event length etc.)
are moved to the SCE's int_colData; forr FACS data, these are now kept inside
the assay data (since FACS channels d+ not contain masses)
differential
+ bug fix in plotFreqHeatmap(): frequencies were previously computed for
the wrong margin (across clusters instead of samples); this has been fixed
+ added checks in prepData() for the existence of md_/panel_cols to give a more
intuitive error message when columns cannot be matched to the input md/panel
+ added flexibility for pbMDS() to allow specification of
- which assay data and summary statistic to use for aggregation
- the level at which to aggregate (cluster, sample, cluster-sample)
+ added vignette section "More - Exporting FCS files" on
how to write FCS files from a SCE using sce2fcs()
changes in version 1.12.0
preprocessing
+ The daFrame class been removed and the preprocessing pipeline
rewritten to use the SingleCellExperiment class instead.
This greatly improves runtimes of debarcoding and normalization,
and affects all functions associated with preprocessing:
- normalization: normCytof()
- debarcoding: assignPrelim(), est/applyCutoffs(), plotEvents/Yields()
- compensation: computeSpillmat(), compCytof(), plotSpillmat()
+ concatFCS() has been replaced by prepData(), which is used during
differential analysis and already provided most of concatFCS()'s
functionality. It has been adapted to
- not require input panel and metadata tables (arguments 'panel' and 'md')
- fix event times in the Time channel as was done previously by concatFCS()
- reorder samples according to their acquisiton time when 'md' is unspecified
+ sce2fcs() has been added to allow for easy conversion from SCE
to flowFrame/Set, which can inturn be written to FCS file(s)
- the SCE can be split into separate frames by a cell metadata variable
(argument 'split_by') to, e.g., write debarcoded samples to separate files
- any cell metadata and dimensionality reductions stored in the SCE can be
optionally included in the output (arguments 'keep_cd' and 'keep_dr')
+ plotScatter() has been added to support basic visualization of biscatters,
but is not meant to compete with other tools available that were designed
specifically for this task (namely, ggcyto)
- cells my be colored by
- density, in which case cells are binned via geom_hex()
- a continuous or categorical cell metadata variable
- facetting is flexible and allows to plot
- one channel against a set of others
- a pair of channels split by tw+ variables
- one channel against others split by one variable
+ plotYields() & plotSpillmat no-longer support interactive plotting with
'plotly'; this did not seem necessary and avoids additional dependencies
+ plotEvents() & plotYields() now support specification of both,
output directory and file name, via arguments 'out_path/name'
differential
+ The metadata()$experiment_inf+ slot is constructed by prepData() and
subsetted by filterSCE() as before. However, it is n+ longer required
for any of the plotting functions associate with differential analysis.
Instead, these rely only on the existence of the colData()$sample_id
(and, for some, colData()$cluster_id) slot(s)
+ Throughout all plotting functions, options for customizing visualizations
(e.g., providing custom color palettes for heatmaps and clusters,
turning on/off row/column clustering and dendrograms in heatmaps etc.)
have been added
+ plotCounts() previously only supported plotting absolute cell counts by
sample and now allows to i) plot relative abundances (frequencies) instead;
and, ii) group cell counts/frequencies by an addition cell metadata variable
(e.g., to plot abundances split by both sample and patient ID)
+ plotMDS() has been replaced by pbMDS() to avoid
Namespace clash with scater::plotMDS()
+ plotDR() failed to visualize PCA results; this has been fixed
+ plotDR() now allows to visualize dimension reductions
colored by an arbitrary number of markers with a single command
+ plotClusterHeatmap() has been deprecated and replaced by
- plotExprHeatmap() for pseudobulk expression heatmaps
(by sample, cluster, cluster-sample)
- plotFreqHeatmap() for relative cluster abundance heatmaps
- plotMultiHeatmap() to combine pseudobulk expression
& cluster frequency heatmaps side-by-side
+ The vignette has been extended to include sections on how
- ggplot2 & ComplexHeatmap visualizations
can be customized or modified in retrospect
- any clustering algorithms other than FlowSOM can be applied and
incorporated to make use of the visualizations available in CATALYST
- arbitrary ComplexHeatmap outputs from plotExprs/FreqHeatmap()
can be combined manually, when not plotMultiHeatmap()
+ An additional vignette has been added demonstrating how CATALYST's
visualizations can be used with other data types from, e.g., scRNA-seq