| Title: | Two-Tier Mapper: a clustering tool based on topological data analysis |
|---|---|
| Description: | TTMap is a clustering method that groups together samples with the same deviation in comparison to a control group. It is specially useful when the data is small. It is parameter free. |
| Authors: | Rachel Jeitziner |
| Maintainer: | Rachel Jeitziner <[email protected]> |
| License: | GPL-2 |
| Version: | 1.29.0 |
| Built: | 2024-11-30 05:35:55 UTC |
| Source: | https://github.com/bioc/TTMap |
TTMap is a clustering method that groups together samples with the same deviation in comparison to a control group. It is specially useful when the data is small. It is parameter free.
The DESCRIPTION file: TTMap/DESCRIPTION Version 1.0
Rachel Jeitziner Maintainer: Rachel Jeitziner <[email protected]>
R. Jeitziner et al., TTMap, 2018, DOI:arXiv:1801.01841
rgl, colorRamps
#to be found in \code{\link[TTMap]{ttmap_sgn_genes}}#to be found in \code{\link[TTMap]{ttmap_sgn_genes}}
Calculation of the value of epsilon
calcul_e(dd5, pvalcutoff = 0.95, tt1, alpha = 1, S = colnames(tt1$Normal.mat)) calcul_e_single(dd5, pvalcutoff = 0.95, tt1, alpha = 1, S = colnames(tt1$Normal.mat))calcul_e(dd5, pvalcutoff = 0.95, tt1, alpha = 1, S = colnames(tt1$Normal.mat)) calcul_e_single(dd5, pvalcutoff = 0.95, tt1, alpha = 1, S = colnames(tt1$Normal.mat))
dd5 |
distance matrix as created by |
pvalcutoff |
cutoff of 0.05 percent (default) or less |
tt1 |
output of |
alpha |
a cutoff value for the FC between the group of control and the disease group |
S |
subset of columns to be considered |
al |
number representing the cutoff to choose for the relatedness with dd5 |
Rachel Jeitziner
control_adjustment,
hyperrectangle_deviation_assessment,
ttmap_sgn_genes,
generate_mismatch_distance
##-- library(airway) data(airway) airway <- airway[rowSums(assay(airway))>80,] assay(airway) <- log(assay(airway)+1,2) ALPHA <- 1 the_experiment <- TTMap::make_matrices(airway, seq_len(4), seq_len(4) + 4, rownames(airway), rownames(airway)) TTMAP_part1prime <-TTMap::control_adjustment( normal.pcl = the_experiment$CTRL, tumor.pcl = the_experiment$TEST, normalname = "The_healthy_controls", dataname = "Effect_of_cancer", org.directory = tempdir(), e = 0, P = 1.1, B = 0); Kprime <- 4; TTMAP_part1_hda <- TTMap::hyperrectangle_deviation_assessment(x = TTMAP_part1prime, k = Kprime,dataname = "Effect_of_cancer", normalname = "The_healthy_controls"); annot <- c(paste(colnames( the_experiment$TEST[,-(seq_len(3))]), "Dis", sep = "."), paste(colnames(the_experiment$CTRL[, -seq_len(3)]), "Dis", sep = ".")) dd5_sgn_only <-TTMap::generate_mismatch_distance( TTMAP_part1_hda, select=rownames(TTMAP_part1_hda$Dc.Dmat), alpha = ALPHA) e <- TTMap::calcul_e(dd5_sgn_only, 0.95, TTMAP_part1prime, 1)##-- library(airway) data(airway) airway <- airway[rowSums(assay(airway))>80,] assay(airway) <- log(assay(airway)+1,2) ALPHA <- 1 the_experiment <- TTMap::make_matrices(airway, seq_len(4), seq_len(4) + 4, rownames(airway), rownames(airway)) TTMAP_part1prime <-TTMap::control_adjustment( normal.pcl = the_experiment$CTRL, tumor.pcl = the_experiment$TEST, normalname = "The_healthy_controls", dataname = "Effect_of_cancer", org.directory = tempdir(), e = 0, P = 1.1, B = 0); Kprime <- 4; TTMAP_part1_hda <- TTMap::hyperrectangle_deviation_assessment(x = TTMAP_part1prime, k = Kprime,dataname = "Effect_of_cancer", normalname = "The_healthy_controls"); annot <- c(paste(colnames( the_experiment$TEST[,-(seq_len(3))]), "Dis", sep = "."), paste(colnames(the_experiment$CTRL[, -seq_len(3)]), "Dis", sep = ".")) dd5_sgn_only <-TTMap::generate_mismatch_distance( TTMAP_part1_hda, select=rownames(TTMAP_part1_hda$Dc.Dmat), alpha = ALPHA) e <- TTMap::calcul_e(dd5_sgn_only, 0.95, TTMAP_part1prime, 1)
control_adjustment function
finds outliers in the control group and
removes them
control_adjustment(normal.pcl, tumor.pcl, normalname, dataname, org.directory = "", A = 1, e = 0, meth = 0, P = 1.1, B = 0)control_adjustment(normal.pcl, tumor.pcl, normalname, dataname, org.directory = "", A = 1, e = 0, meth = 0, P = 1.1, B = 0)
normal.pcl |
the control matrix with annotation as obtained by $CTRL from
|
tumor.pcl |
the disease/test data matrix with annotation as obtained by $TEST
from |
normalname |
A name for the corrected control files |
dataname |
the name of the project |
org.directory |
where the outputs should be saved |
A |
integer if A=0 then the difference to the median is calculated otherwise the difference to the mean. |
e |
integer giving how far to the median an outlier is at least |
meth |
value or method that defines how to replace outliers, default is set to replace by the median |
P |
if more than P percent of features are outliers the feature is removed, by default all are kept |
B |
Batch vector a vector for normal and test samples with a same number corresponding to a same batch |
control_adjustment calculates a
corrected control group, discovers
outliers in it.
Several files are created
paste(org.directory, normalname, ".normMesh", sep = "")
|
The normal matrix with only common features with the test matrix. This file is only created if the two have different rows |
paste(org.directory, dataname, ".normMesh", sep = "")
|
The test matrix with only common features with the normal matrix. This file is only created if the two have different rows. |
mean_vs_variance.pdf |
A pdf showing a plot of the mean (X axis) against the variances (Y axis) of each feature |
mean_vs_variance_after_correction.pdf |
A pdf showing a plot of the mean (X axis) against the variances (Y axis) of each feature after correction of the control group |
na_numbers_per_row.txt |
number of outliers per row |
na_numbers_per_col.txt |
number of outliers per column |
And values of ttmap_part1_ctrl_adj
e |
Selected criteria for what is an outlier |
tag.pcl |
Annotation of features, ID of features and weight |
Normal.mat |
The control matrix without annotation and only with the common rows with Disease.mat |
Disease.mat |
The test/disease matrix without annotation and only with the common rows with Disease.mat |
flat.Nmat |
A list $mat being the corrected control matrix $m a record of the different numbers of removed genes per sample |
record |
numbers recording the number of columns in Disease.mat and Normal.mat |
B |
The batch vector B introduced in the begining |
U1 |
The different batches in Normal.mat |
U2 |
The different batches in Disease.mat |
Rachel Jeitziner
hyperrectangle_deviation_assessment,
ttmap
ttmap_sgn_genes
##-- library(airway) data(airway) airway <- airway[rowSums(assay(airway))>80,] assay(airway) <- log(assay(airway)+1,2) ALPHA <- 1 the_experiment <- TTMap::make_matrices(airway, seq_len(4), seq_len(4) + 4, rownames(airway), rownames(airway)) TTMAP_part1prime <-TTMap::control_adjustment( normal.pcl = the_experiment$CTRL, tumor.pcl = the_experiment$TEST, normalname = "The_healthy_controls", dataname = "Effect_of_cancer", org.directory = tempdir(), e = 0, P = 1.1, B = 0);##-- library(airway) data(airway) airway <- airway[rowSums(assay(airway))>80,] assay(airway) <- log(assay(airway)+1,2) ALPHA <- 1 the_experiment <- TTMap::make_matrices(airway, seq_len(4), seq_len(4) + 4, rownames(airway), rownames(airway)) TTMAP_part1prime <-TTMap::control_adjustment( normal.pcl = the_experiment$CTRL, tumor.pcl = the_experiment$TEST, normalname = "The_healthy_controls", dataname = "Effect_of_cancer", org.directory = tempdir(), e = 0, P = 1.1, B = 0);
Single cell complete mismatch distance, single cell complete mismatch distance with a parameter of cutoff, mismatch distance, correlation distance, p-value of correlation test distance and euclidean distance.
generate_single_cell_complete_mismatch(ttmap_part1_hda, select, alpha = 1) generate_single_cell_mismatch_with_parameter(ttmap_part1_hda, select, alpha = 1) generate_correlation(ttmap_part1_hda, select) generate_euclidean(ttmap_part1_hda, select) generate_mismatch_distance(ttmap_part1_hda, select, alpha = 1) generate_p_val_correlation(ttmap_part1_hda, select)generate_single_cell_complete_mismatch(ttmap_part1_hda, select, alpha = 1) generate_single_cell_mismatch_with_parameter(ttmap_part1_hda, select, alpha = 1) generate_correlation(ttmap_part1_hda, select) generate_euclidean(ttmap_part1_hda, select) generate_mismatch_distance(ttmap_part1_hda, select, alpha = 1) generate_p_val_correlation(ttmap_part1_hda, select)
ttmap_part1_hda |
an object given back
by |
select |
A sublist of rownames of ttmap_part1_hda$Dc.Dmat |
alpha |
A real number corresponding to a cutoff |
If one is interested only in clustering samples according to a list of genes belonging to a certain pathway, then this list is provided to the parameter select. Alpha is a cutoff for deviations that should be considered as noise, for gene expression data such as normalised RNA-seq or microarrays for instance a cutoff of 1, corresponding to a two fold change is being chosen.
Distance matrix
Rachel Jeitziner
ttmap_part1_hda <- list() ttmap_part1_hda$Dc.Dmat <- matrix(c(-1, 2, 0, -4, 5, 6), nrow = 2) rownames(ttmap_part1_hda$Dc.Dmat) <- c("Gene1", "Gene2") colnames(ttmap_part1_hda$Dc.Dmat) <- c("A", "B", "C") dd <- TTMap::generate_mismatch_distance(ttmap_part1_hda, select = rownames(ttmap_part1_hda$Dc.Dmat)) dd <- TTMap::generate_euclidean(ttmap_part1_hda, select = rownames(ttmap_part1_hda$Dc.Dmat))ttmap_part1_hda <- list() ttmap_part1_hda$Dc.Dmat <- matrix(c(-1, 2, 0, -4, 5, 6), nrow = 2) rownames(ttmap_part1_hda$Dc.Dmat) <- c("Gene1", "Gene2") colnames(ttmap_part1_hda$Dc.Dmat) <- c("A", "B", "C") dd <- TTMap::generate_mismatch_distance(ttmap_part1_hda, select = rownames(ttmap_part1_hda$Dc.Dmat)) dd <- TTMap::generate_euclidean(ttmap_part1_hda, select = rownames(ttmap_part1_hda$Dc.Dmat))
hyperrectangle_deviation_assessment function
calculates the
hyperrectangle deviation
assessment (HDA) that calculates the
deviation components using
normal_hda2 which calculates
the normal component of the test
sample and deviation_hda2
which calculates the deviation component.
hyperrectangle_deviation_assessment(x, k = dim(x$Normal.mat)[2], dataname, normalname,Org.directory = getwd())hyperrectangle_deviation_assessment(x, k = dim(x$Normal.mat)[2], dataname, normalname,Org.directory = getwd())
x |
output object given back
by |
k |
A factor if not all the lines in the control group should be kept |
dataname |
the name of the project |
normalname |
A name for the corrected control files |
Org.directory |
where the outputs should be saved |
The function performs the hyperrectangle deviation assessment (HDA)
Outputs
Tdis.pcl |
The matrix of the deviation components for each test sample |
Tnorm.pcl |
The matrix of the normal components for each test sample |
NormalModel.pcl |
The normal model used |
Values
Dc.Dmat |
the deviation component matrix composed of the deviation components of all the samples in the test group |
m |
the values of the filter function per sample in the test group |
Rachel Jeitziner
control_adjustment,
hyperrectangle_deviation_assessment,
ttmap_sgn_genes
##a full example can be found in ttmap_sgn_genes ##-- library(airway) data(airway) airway <- airway[rowSums(assay(airway))>80,] assay(airway) <- log(assay(airway)+1,2) ALPHA <- 1 the_experiment <- TTMap::make_matrices(airway, seq_len(4), seq_len(4) + 4, rownames(airway), rownames(airway)) TTMAP_part1prime <-TTMap::control_adjustment( normal.pcl = the_experiment$CTRL, tumor.pcl = the_experiment$TEST, normalname = "The_healthy_controls", dataname = "Effect_of_cancer", org.directory = tempdir(), e = 0, P = 1.1, B = 0); Kprime <- 4; TTMAP_part1_hda <- TTMap::hyperrectangle_deviation_assessment(x = TTMAP_part1prime, k = Kprime, dataname = "Effect_of_cancer", normalname = "The_healthy_controls");##a full example can be found in ttmap_sgn_genes ##-- library(airway) data(airway) airway <- airway[rowSums(assay(airway))>80,] assay(airway) <- log(assay(airway)+1,2) ALPHA <- 1 the_experiment <- TTMap::make_matrices(airway, seq_len(4), seq_len(4) + 4, rownames(airway), rownames(airway)) TTMAP_part1prime <-TTMap::control_adjustment( normal.pcl = the_experiment$CTRL, tumor.pcl = the_experiment$TEST, normalname = "The_healthy_controls", dataname = "Effect_of_cancer", org.directory = tempdir(), e = 0, P = 1.1, B = 0); Kprime <- 4; TTMAP_part1_hda <- TTMap::hyperrectangle_deviation_assessment(x = TTMAP_part1prime, k = Kprime, dataname = "Effect_of_cancer", normalname = "The_healthy_controls");
control_adjustment
make_matrices generates the control
and the test matrice in the
right format
make_matrices(mat, col_ctrl, col_test, NAME, CLID, GWEIGHT = rep(1, dim(mat)[1]), EWEIGHT = 0)make_matrices(mat, col_ctrl, col_test, NAME, CLID, GWEIGHT = rep(1, dim(mat)[1]), EWEIGHT = 0)
mat |
the gene expressions can be matrix, data.frame,
|
col_ctrl |
the columns in the matrix "mat" of the control samples |
col_test |
the columns in the matrix "mat" of the test samples |
NAME |
Name of genes,or annotation, e.g. WNT4 |
CLID |
Identities of genes,e.g. ENSMUSG00000000001 |
GWEIGHT |
the weight for each gene |
EWEIGHT |
the weight for each experiment |
make_matrices generates the test matrix
and the control matrix in
the format accepted by
control_adjustment
from a matrix object
junk |
A list containing $CTRL and $TEST the matrices
to impute
in |
Rachel Jeitziner
control_adjustment,
hyperrectangle_deviation_assessment,
ttmap_sgn_genes,
"
RangedSummarizedExperiment"
##-- ##-- Aa = 6 B1 = 3 B2 = 3 C0 = 100 D0 = 10000 a0 = 4 b0 = 0.1 a1 = 6 b1 = 0.1 a2 = 2 b2 = 0.5 ALPHA = 1 E = 1 Pw = 1.1 Bw = 0 RA <- matrix(rep(0, Aa * D0), nrow = D0) RB1 <- matrix(rep(0, B1 * D0), nrow = D0) RB2 <- matrix(rep(0, B2 * D0), nrow = D0) RA <- lapply(seq_len(D0 - C0), function(i) rnorm(Aa, mean = a0, sd = sqrt(b0))) RA<-do.call(rbind, RA) RB1<- lapply(seq_len(D0 - C0), function(i) rnorm(B1, mean = a0, sd = sqrt(b0))) RB1 <- do.call(rbind, RB1) RB2 <- lapply(seq_len(D0 - C0), function(i) rnorm(B2, mean = a0, sd = sqrt(b0))) RB2 <- do.call(rbind, RB2) RA_c <- lapply(seq_len(C0), function(i) rnorm(Aa, mean = a0, sd = sqrt(b0))) RA_c <- do.call(rbind, RA_c) RB1_c <- lapply(seq_len(C0), function(i) rnorm(B1, mean = a1, sd = sqrt(b1))) RB1_c <- do.call(rbind, RB1_c) RB2_c <- lapply(seq_len(C0), function(i) rnorm(B2, mean = a2, sd = sqrt(b2))) RB2_c <- do.call(rbind, RB2_c) norm1 <- rbind(RA, RA_c) dis <- cbind(rbind(RB1, RB1_c), rbind(RB2, RB2_c)) colnames(norm1) <- paste("N", seq_len(Aa), sep = "") rownames(norm1) <- c(paste("norm", seq_len(D0 - C0), sep = ""), paste("diff", seq_len(C0), sep = "")) colnames(dis) <- c(paste("B1", seq_len(B1), sep=""), paste("B2", seq_len(B2), sep ="")) rownames(dis)<-c(paste("norm", seq_len(D0 - C0), sep = ""), paste("diff", seq_len(C0), sep = "")) the_experiment <- TTMap::make_matrices(cbind(norm1, dis), col_ctrl = colnames(norm1), col_test = colnames(dis), NAME = rownames(norm1), CLID = rownames(norm1)) ###other example using SummarizedExperiment library(airway) data(airway) airway <- airway[rowSums(assay(airway))>80,] assay(airway) <- log(assay(airway)+1,2) the_experiment <- TTMap::make_matrices(airway, seq_len(4), seq_len(4) + 4, rownames(airway), rownames(airway))##-- ##-- Aa = 6 B1 = 3 B2 = 3 C0 = 100 D0 = 10000 a0 = 4 b0 = 0.1 a1 = 6 b1 = 0.1 a2 = 2 b2 = 0.5 ALPHA = 1 E = 1 Pw = 1.1 Bw = 0 RA <- matrix(rep(0, Aa * D0), nrow = D0) RB1 <- matrix(rep(0, B1 * D0), nrow = D0) RB2 <- matrix(rep(0, B2 * D0), nrow = D0) RA <- lapply(seq_len(D0 - C0), function(i) rnorm(Aa, mean = a0, sd = sqrt(b0))) RA<-do.call(rbind, RA) RB1<- lapply(seq_len(D0 - C0), function(i) rnorm(B1, mean = a0, sd = sqrt(b0))) RB1 <- do.call(rbind, RB1) RB2 <- lapply(seq_len(D0 - C0), function(i) rnorm(B2, mean = a0, sd = sqrt(b0))) RB2 <- do.call(rbind, RB2) RA_c <- lapply(seq_len(C0), function(i) rnorm(Aa, mean = a0, sd = sqrt(b0))) RA_c <- do.call(rbind, RA_c) RB1_c <- lapply(seq_len(C0), function(i) rnorm(B1, mean = a1, sd = sqrt(b1))) RB1_c <- do.call(rbind, RB1_c) RB2_c <- lapply(seq_len(C0), function(i) rnorm(B2, mean = a2, sd = sqrt(b2))) RB2_c <- do.call(rbind, RB2_c) norm1 <- rbind(RA, RA_c) dis <- cbind(rbind(RB1, RB1_c), rbind(RB2, RB2_c)) colnames(norm1) <- paste("N", seq_len(Aa), sep = "") rownames(norm1) <- c(paste("norm", seq_len(D0 - C0), sep = ""), paste("diff", seq_len(C0), sep = "")) colnames(dis) <- c(paste("B1", seq_len(B1), sep=""), paste("B2", seq_len(B2), sep ="")) rownames(dis)<-c(paste("norm", seq_len(D0 - C0), sep = ""), paste("diff", seq_len(C0), sep = "")) the_experiment <- TTMap::make_matrices(cbind(norm1, dis), col_ctrl = colnames(norm1), col_test = colnames(dis), NAME = rownames(norm1), CLID = rownames(norm1)) ###other example using SummarizedExperiment library(airway) data(airway) airway <- airway[rowSums(assay(airway))>80,] assay(airway) <- log(assay(airway)+1,2) the_experiment <- TTMap::make_matrices(airway, seq_len(4), seq_len(4) + 4, rownames(airway), rownames(airway))
control_adjustment
make_matrices generates the control (output $CTRL)
and the test (output $TEST) matrice in the
right format for control_adjustment
signature(mat = "data.frame")Method make_matrice
for data.frame object.
signature(mat = "matrix")Method make_matrice
for matrix object.
signature(mat = "SummarizedExperiment")Method make_matrice
for SummarizedExperiment object.
signature(mat = "RangedSummarizedExperiment")Method make_matrice
for RangedSummarizedExperiment object.
signature(mat = "ExpressionSet")Method make_matrice
for ExpressionSet object.
Enables a quick view on the groups in the dataset (globally) and how locally they differ.
ttmap(ttmap_part1_hda, m1, select = row.names(ttmap_part1_hda$Dc.Dmat), ddd, e, filename = "TEST", n = 3, ad = 0, bd = 0, piq = 1, dd = generate_mismatch_distance(ttmap_part1_hda = ttmap_part1_hda, select = select), mean_value_m1 = "N", ni = 2)ttmap(ttmap_part1_hda, m1, select = row.names(ttmap_part1_hda$Dc.Dmat), ddd, e, filename = "TEST", n = 3, ad = 0, bd = 0, piq = 1, dd = generate_mismatch_distance(ttmap_part1_hda = ttmap_part1_hda, select = select), mean_value_m1 = "N", ni = 2)
ttmap_part1_hda |
list output of |
m1 |
either a user imputed vector whose names are the names of the samples with addition of .Dis. or by default it is the amount of deviation |
select |
Should all the features (default) or only a sublist be considered to calculate the distance |
ddd |
Annotation matrix with rownames the different sample names with addition of .Dis. There can be as many columns as wanted, but only the column n will be selected to annotated the clusters |
e |
integer parameter defining under which value two samples are considered to be close |
filename |
Name for the description file annotating the clusters |
n |
The column to be considered to annotate the clusters |
ad |
if ad!=0 then the clusters on the output picture will not be annotated |
bd |
if different than 0 (default), the output will be without outliers of the test data set (clusters composed of only "piq" element) |
piq |
parameter used to determine what small clusters are, see bd |
dd |
the distance matrix to be used |
mean_value_m1 |
if == "N" the average of the values in m1 divided by the number of the samples are put into the legend (by default represents the average of the samples in a cluster of the mean-deviation of the features) otherwise it will show the average value of the values in m1 (is useful for instance if m1 represents the age of the samples) |
ni |
The column to consider to annotate the samples (is put into parenthesis) for the description file |
Is the Two-tiers Mapper function. The output is an interactive image of the clusters in the different layers.
all |
the clusters in the overall group |
low |
the clusters in the lower quartile group |
mid1 |
the clusters in the first middle quartile group |
mid2 |
the clusters in the second middle quartile group |
high |
the clusters in the higher quartile group |
Rachel Jeitziner
control_adjustment,
hyperrectangle_deviation_assessment,
ttmap_sgn_genes
##-- library(airway) data(airway) airway <- airway[rowSums(assay(airway))>80,] assay(airway) <- log(assay(airway)+1,2) ALPHA <- 1 the_experiment <- TTMap::make_matrices(airway, seq_len(4), seq_len(4) + 4, rownames(airway), rownames(airway)) TTMAP_part1prime <-TTMap::control_adjustment( normal.pcl = the_experiment$CTRL, tumor.pcl = the_experiment$TEST, normalname = "The_healthy_controls", dataname = "Effect_of_cancer", org.directory = tempdir(), e = 0, P = 1.1, B = 0); Kprime <- 4; TTMAP_part1_hda <- TTMap::hyperrectangle_deviation_assessment(x = TTMAP_part1prime, k = Kprime,dataname = "Effect_of_cancer", normalname = "The_healthy_controls"); annot <- c(paste(colnames( the_experiment$TEST[,-(seq_len(3))]),"Dis", sep = "."), paste(colnames(the_experiment$CTRL[, -seq_len(3)]), "Dis", sep = ".")) annot <- cbind(annot, annot) rownames(annot)<-annot[, 1] dd5_sgn_only <-TTMap::generate_mismatch_distance( TTMAP_part1_hda, select=rownames(TTMAP_part1_hda$Dc.Dmat), alpha = ALPHA) TTMAP_part2 <- TTMap::ttmap(TTMAP_part1_hda, TTMAP_part1_hda$m, select = rownames(TTMAP_part1_hda$Dc.Dmat), annot, e = TTMap::calcul_e(dd5_sgn_only, 0.95, TTMAP_part1prime, 1), filename = "first_comparison", n = 1, dd = dd5_sgn_only)##-- library(airway) data(airway) airway <- airway[rowSums(assay(airway))>80,] assay(airway) <- log(assay(airway)+1,2) ALPHA <- 1 the_experiment <- TTMap::make_matrices(airway, seq_len(4), seq_len(4) + 4, rownames(airway), rownames(airway)) TTMAP_part1prime <-TTMap::control_adjustment( normal.pcl = the_experiment$CTRL, tumor.pcl = the_experiment$TEST, normalname = "The_healthy_controls", dataname = "Effect_of_cancer", org.directory = tempdir(), e = 0, P = 1.1, B = 0); Kprime <- 4; TTMAP_part1_hda <- TTMap::hyperrectangle_deviation_assessment(x = TTMAP_part1prime, k = Kprime,dataname = "Effect_of_cancer", normalname = "The_healthy_controls"); annot <- c(paste(colnames( the_experiment$TEST[,-(seq_len(3))]),"Dis", sep = "."), paste(colnames(the_experiment$CTRL[, -seq_len(3)]), "Dis", sep = ".")) annot <- cbind(annot, annot) rownames(annot)<-annot[, 1] dd5_sgn_only <-TTMap::generate_mismatch_distance( TTMAP_part1_hda, select=rownames(TTMAP_part1_hda$Dc.Dmat), alpha = ALPHA) TTMAP_part2 <- TTMap::ttmap(TTMAP_part1_hda, TTMAP_part1_hda$m, select = rownames(TTMAP_part1_hda$Dc.Dmat), annot, e = TTMap::calcul_e(dd5_sgn_only, 0.95, TTMAP_part1prime, 1), filename = "first_comparison", n = 1, dd = dd5_sgn_only)
ttmap_sgn_genes function
ttmap_sgn_genes(ttmap_part2_gtlmap, ttmap_part1_hda, ttmap_part1_ctrl_adj, c, n = 2, a = 0, filename = "TEST2", annot = ttmap_part1_ctrl_adj$tag.pcl, col = "NAME", path = getwd(), Relaxed = 1) ttmap_sgn_genes_inter2(q, ttmap_part1_hda, alpha = 0) ttmap_sgn_genes_inter(q, ttmap_part1_hda, alpha = 0)ttmap_sgn_genes(ttmap_part2_gtlmap, ttmap_part1_hda, ttmap_part1_ctrl_adj, c, n = 2, a = 0, filename = "TEST2", annot = ttmap_part1_ctrl_adj$tag.pcl, col = "NAME", path = getwd(), Relaxed = 1) ttmap_sgn_genes_inter2(q, ttmap_part1_hda, alpha = 0) ttmap_sgn_genes_inter(q, ttmap_part1_hda, alpha = 0)
ttmap_part2_gtlmap |
output of |
ttmap_part1_hda |
output of |
ttmap_part1_ctrl_adj |
output of |
c |
annotation file of the samples |
n |
column to give the name to the cluster |
a |
cutoff to be considered different than noise |
filename |
Name of the files |
annot |
annotation file |
col |
which column should be considered to annotate the features |
path |
where to put the output files |
Relaxed |
If Relaxed then one allows sample to be as the control and for all the others in one cluster to be going in the same direction (more than alpha) otherwise all the features must be deviating to be considered a significant feature |
q |
The sample in one cluster |
alpha |
cutoff to be considered different than noise inherited by a |
Is giving per cluster the features that vary in the same direction
generates a file per cluster of significant features with an annotation
Rachel Jeitziner
##-- library(airway) data(airway) airway <- airway[rowSums(assay(airway))>80,] assay(airway) <- log(assay(airway)+1,2) ALPHA <- 1 the_experiment <- TTMap::make_matrices(airway, seq_len(4), seq_len(4) + 4, rownames(airway), rownames(airway)) TTMAP_part1prime <-TTMap::control_adjustment( normal.pcl = the_experiment$CTRL, tumor.pcl = the_experiment$TEST, normalname = "The_healthy_controls", dataname = "Effect_of_cancer", org.directory = tempdir(), e = 0, P = 1.1, B = 0); Kprime <- 4; TTMAP_part1_hda <- TTMap::hyperrectangle_deviation_assessment(x = TTMAP_part1prime, k = Kprime,dataname = "Effect_of_cancer", normalname = "The_healthy_controls"); annot <- c(paste(colnames( the_experiment$TEST[,-(seq_len(3))]),"Dis", sep = "."), paste(colnames(the_experiment$CTRL[, -seq_len(3)]), "Dis", sep = ".")) annot <- cbind(annot, annot) rownames(annot)<-annot[, 1] dd5_sgn_only <-TTMap::generate_mismatch_distance( TTMAP_part1_hda, select=rownames(TTMAP_part1_hda$Dc.Dmat), alpha = ALPHA) TTMAP_part2 <- TTMap::ttmap(TTMAP_part1_hda, TTMAP_part1_hda$m, select = rownames(TTMAP_part1_hda$Dc.Dmat), annot, e = TTMap::calcul_e(dd5_sgn_only, 0.95, TTMAP_part1prime, 1), filename = "first_comparison", n = 1, dd = dd5_sgn_only) TTMap::ttmap_sgn_genes(TTMAP_part2, TTMAP_part1_hda, TTMAP_part1prime, annot, n = 2, a = 1, filename = "first_list_of_genes", annot = TTMAP_part1prime$tag.pcl, col = "NAME", path = getwd(), Relaxed = 1)##-- library(airway) data(airway) airway <- airway[rowSums(assay(airway))>80,] assay(airway) <- log(assay(airway)+1,2) ALPHA <- 1 the_experiment <- TTMap::make_matrices(airway, seq_len(4), seq_len(4) + 4, rownames(airway), rownames(airway)) TTMAP_part1prime <-TTMap::control_adjustment( normal.pcl = the_experiment$CTRL, tumor.pcl = the_experiment$TEST, normalname = "The_healthy_controls", dataname = "Effect_of_cancer", org.directory = tempdir(), e = 0, P = 1.1, B = 0); Kprime <- 4; TTMAP_part1_hda <- TTMap::hyperrectangle_deviation_assessment(x = TTMAP_part1prime, k = Kprime,dataname = "Effect_of_cancer", normalname = "The_healthy_controls"); annot <- c(paste(colnames( the_experiment$TEST[,-(seq_len(3))]),"Dis", sep = "."), paste(colnames(the_experiment$CTRL[, -seq_len(3)]), "Dis", sep = ".")) annot <- cbind(annot, annot) rownames(annot)<-annot[, 1] dd5_sgn_only <-TTMap::generate_mismatch_distance( TTMAP_part1_hda, select=rownames(TTMAP_part1_hda$Dc.Dmat), alpha = ALPHA) TTMAP_part2 <- TTMap::ttmap(TTMAP_part1_hda, TTMAP_part1_hda$m, select = rownames(TTMAP_part1_hda$Dc.Dmat), annot, e = TTMap::calcul_e(dd5_sgn_only, 0.95, TTMAP_part1prime, 1), filename = "first_comparison", n = 1, dd = dd5_sgn_only) TTMap::ttmap_sgn_genes(TTMAP_part2, TTMAP_part1_hda, TTMAP_part1prime, annot, n = 2, a = 1, filename = "first_list_of_genes", annot = TTMAP_part1prime$tag.pcl, col = "NAME", path = getwd(), Relaxed = 1)
Reading (read_pcl),
writing (write_pcl)
files and annotating matrices (mat2pcl)
mat2pcl(mat, tag) write_pcl(df, dataname, fileaddress = "") read_pcl(filename, na.type = "", Nrows = -1, Comment.char = "", ...)mat2pcl(mat, tag) write_pcl(df, dataname, fileaddress = "") read_pcl(filename, na.type = "", Nrows = -1, Comment.char = "", ...)
df |
PCL object to be saved |
dataname |
Name of the file |
fileaddress |
Where to save the file |
filename |
File name to be loaded on R |
na.type |
feels the parameter na.strings of read.table |
Nrows |
Number of rows to be ignored (nrows of read.table) |
Comment.char |
comment.char of read.table |
... |
other read.table arguments |
mat |
matrix to be changed in annotated |
tag |
annotation |
The file (called filename) MUST contain 3 columns before the actual values, which are called CLID, NAME and GWEIGHT, described bellow. The first row must be the header of the columns (starting with CLID,NAME and GWEIGHT) and the second row must be EWEIGHT. Representing how much weight each column has: if some columns are n replicates they can have each a weight of 1/n.
Data frame composed of
CLID |
Column called CLID which is the ID of the features, which will then be the rownames of the dataframe |
NAME |
A possibly longer name, more meaningfull than CLID, text format |
GWEIGHT |
A weight for each gene or feature. If some genes are less important than others or only a pathway should be selected than the file (called filename) should have this information |
Matrix |
The matrix with numbers of the different observations |
Rachel Jeitziner
library(airway) data(airway) airway <- airway[rowSums(assay(airway))>80,] assay(airway) <- log(assay(airway)+1,2) ALPHA <- 1 to_be_saved <- TTMap::make_matrices(airway, seq_len(4), seq_len(4) + 4, rownames(airway), rownames(airway)) TTMap::write_pcl(to_be_saved, "tempfile()", getwd())library(airway) data(airway) airway <- airway[rowSums(assay(airway))>80,] assay(airway) <- log(assay(airway)+1,2) ALPHA <- 1 to_be_saved <- TTMap::make_matrices(airway, seq_len(4), seq_len(4) + 4, rownames(airway), rownames(airway)) TTMap::write_pcl(to_be_saved, "tempfile()", getwd())