"plot"(x, threshold = 0.05, myproc = "BY", nbsimu_pval = 1e+06, expr, Subject_ID, TimePoint, baseline = NULL, only.signif = TRUE, group.var = NULL, Group_ID_paired = NULL, ref = NULL, group_of_interest = NULL, ranking = FALSE, FUNcluster = NULL, clustering_metric = "euclidian", clustering_method = "ward", B = 500, max_trends = 4, aggreg.fun = "median", methodOptiClust = "firstSEmax", indiv = "genes", verbose = TRUE, clust_trends = NULL, N_clusters = NULL, myclusters = NULL, label.clusters = NULL, prev_rowCL = NULL, descript = TRUE, plot = TRUE, color.vec = c("darkred", "#D73027", "#FC8D59", "snow", "#91BFDB", "#4575B4", "darkblue"), legend.breaks = NULL, label.column = NULL, time_unit = "", cex.label.row = 1, cex.label.column = 1, margins = c(5, 25), heatKey.size = 1, dendrogram.size = 1, heatmap.height = 1, heatmap.width = 1, cex.clusterKey = 1, cex.main = 1, horiz.clusterKey = TRUE, main = NULL, subtitle = NULL, ...)TcGSA'.Bonferroni",
"Holm", "Hochberg", "SidakSS", "SidakSD",
"BH", "BY", "ABH", "TSBH" or "none".
"none" indicates no adjustement for multiple testing. See
mt.rawp2adjp for details. Default is
"BY", the Benjamini & Yekutieli (2001) step-up FDR-controlling
procedure (general dependency structures). In order to control the FWER(in
case of an analysis that is more a hypothesis confirmation than an
exploration of the expression data), we recommand to use "Holm", the
Holm (1979) step-down adjusted p-values for strong control of the FWER.1e+06.'Estimations' element of TcGSA.LR. See details.expr (when it is a dataframe) and that contains the patient
identifier of each sample. Ignored if expr is a list of estimations.Subject_ID and the columns of expr (when it
is a dataframe), and that contains the time points at which gene expression
was measured. Ignored if expr is a list of estimations.TimePoint
used as baseline. See Details.FALSE, all the gene sets from the gmt object contained in
x are plotted. Default is TRUE.Timepoint,
Subject_ID, sample_name and the columns of expr. It
indicates to which treatment group each sample belongs to. Default is
NULL, which means that there is only one treatment group. See
Details.Timepoint, Subject_ID, sample_name,
group.var and the columns of expr. This argument must not be
NULL in the case of a paired analysis, and must be NULL
otherwise. Default is NULL. See Details.NULL, which means that reference is the
first group in alphabetical order of the labels of group.var. See
Details.NULL,
which means that group of interest is the second group in alphabetical order
of the labels of group.var. See Details.
group_of_interest here~~TRUE, the gene set trends are not hierarchicaly classified, but
ordered by decreasing Likelihood ratios. Default is FALSE.FUNcluster is NULL. The currently available
options are "euclidean" and "manhattan". Default is
"euclidean". See agnes. Also, a "sts" option
is available in TcGSA. It implements the 'Short Time Series' distance
[Moller-Levet et al., Fuzzy CLustering of short time series and unevenly distributed
sampling points, Advances in Intelligent Data Analysis V:330-340 Springer, 2003]
designed specifically for clustering time series.FUNcluster is
NULL. The six methods implemented are "average" ([unweighted
pair-]group average method, UPGMA), "single" (single linkage),
"complete" (complete linkage), "ward" (Ward's method),
"weighted" (weighted average linkage). Default is "ward". See
agnes.500. See
clusGap.4."mean", "median"
or the name of any other statistics function defined that returns a single
numeric value. It specifies the function used to aggregate the observations
before the clustering. Default is to median. Default is
"median"."globalmax", "firstmax",
"Tibs2001SEmax", "firstSEmax" and "globalSEmax".
Default is "firstSEmax". See 'method' in
clusGap, Details and Tibshirani et al.,
2001 in References.aggreg.fun) before the clustering. Possible
values are "genes" or "patients". Default is "genes".TRUE.NULL.NULL, in which case
the dendrogram is not cut and no clusters are identified.N_clusters.
Default is NULL, in which case the clusters are automatically
identified and colored via the cutree function and the
N_clusters argument only.N_clusters is not NULL, a character
vector of length N_clusterss. Default is NULL, in which case
if N_clusters is not NULL, clusters are simply labelled with
numbers.NULL, no
clustering is calculated by the present plotting function and this tree is
used to represent the gene sets dynamics. Default is NULL.TRUE.
Default is TRUE. See Details.TRUE.c("#D73027", "#FC8D59","lightyellow", "#91BFDB", "#4575B4").NULL, in which case the break points will be
spaced equally and symetrically about 0.NULL."Y",
"M", "W", "D", "H", etc) next to the values of
TimePoint in the columns labels when label.column is
NULL. Default is "".1.1.par(mar= *)) for column and row names, respectively. Default
is c(15, 100). See Details.1.1111,
when N_clusters is not NULL.1.TRUE, set the legend for
clusters horizontally rather than vertically. Only used if the
N_clusters argument is not NULL. Default is TRUE.NULL.NULL.merge an $n-1$ by $2$ matrix. Row $i$ of
merge describes the merging of clusters at step i of the clustering.
If an element $j$ in the row is negative, then observation -$j$ was
merged at this stage. If $j$ is positive then the merge was with the
cluster formed at the (earlier) stage $j$ of the algorithm. Thus
negative entries in merge indicate agglomerations of singletons, and positive
entries indicate agglomerations of non-singletons.
height a set of $n-1$ real values (non-decreasing for
ultrametric trees). The clustering height: that is, the value of the
criterion associated with the Ward clustering method.
order a vector giving the permutation of the original
observations suitable for plotting, in the sense that a cluster plot using
this ordering and matrix merge will not have crossings of the branches.
labels the gene set trends name.
call the call which produced the result clustering:
hclust(d = dist(map2heat, method = "euclidean"), method = "ward.D2")
method "ward.D2", as it is the clustering method that has been used
for clustering the gene set trends.
dist.method "euclidean", as it is the distance that has been used
for clustering the gene set trends.
legend.breaks a numeric vector giving the splitting points used
for coloring the heatmap. If plot is FALSE, then it is
NULL.
myclusters a character vector of colors for the dynamic clusters
of the represented gene set trends, with as many levels as the value of
N_clusters. If no dynamic clusters were represented, than this is
NULL.
ddr a dendrogram object with the reordering used for the
heatmap. See heatmap.2.
clust.trends a ClusteredTrends object.
clustersExport a data frame with 2 variables containing the two
following variables : GeneSet: the gene set trends
clustered. Cluster: the dynamic cluster they belong to. Cluster.
data_plotted: the data matrix represented by the heatmap
If expr is a matrix or a dataframe, then the "original" data are
plotted. On the other hand, if expr is a list returned in the
'Estimations' element of TcGSA.LR, then it is those
"estimations" made by the TcGSA.LR function that are plotted.
If descript is FALSE, the second element of margins can
be reduced (for instance use margins = c(5, 10)), as there is not so
much need for space in order to display only the gene set names, without
their description.
If there is a large number of significant gene sets, the hierarchical clustering
step repeated for each of them can take a few minutes. To speed things up
(especially) when playing with the ploting parameters for having a nice plot,
one can run the clustTrend function beforehand, and plug its results
in the plot.TcGSA function via the clust_trends argument.
heatmap.2, TcGSA.LR,
hclust
data(data_simu_TcGSA)
tcgsa_sim_1grp <- TcGSA.LR(expr=expr_1grp, gmt=gmt_sim, design=design,
subject_name="Patient_ID", time_name="TimePoint",
time_func="linear", crossedRandom=FALSE)
summary(tcgsa_sim_1grp)
plot(x=tcgsa_sim_1grp, expr=tcgsa_sim_1grp$Estimations,
Subject_ID=design$Patient_ID, TimePoint=design$TimePoint,
baseline=1,
B=100,
time_unit="H",
dendrogram.size=0.4, heatmap.width=0.8, heatmap.height=2, cex.main=0.7
)
## Not run:
# tcgsa_sim_2grp <- TcGSA.LR(expr=expr_2grp, gmt=gmt_sim, design=design,
# subject_name="Patient_ID", time_name="TimePoint",
# time_func="linear", crossedRandom=FALSE,
# group_name="group.var")
# summary(tcgsa_sim_2grp)
# plot(x=tcgsa_sim_2grp, expr=expr_2grp,
# Subject_ID=design$Patient_ID, TimePoint=design$TimePoint,
# B=100,
# time_unit="H",
# )
# ## End(Not run)
Run the code above in your browser using DataLab