Package 'uem915'

Principal Component Analysis

Description

Principal Component Analysis

Usage

acp(
  dat,
  factor = NULL,
  samplename = NULL,
  pc1 = 1,
  pc2 = 2,
  center = TRUE,
  scale = TRUE,
  title = "ACP",
  legendtitle = "TREATMENT"
)

Arguments

dat	matrix numeric
factor	factor
samplename	character
pc1	numeric
pc2	numeric
center	logical TRUE
scale	logical TRUE
title	character
legendtitle	character

Value

no values

Author(s)

Florent Dumont [email protected]

Examples

# not run
# acp( mat1 , sif1 )

---

Annotation

Description

Annotate a list of symbols or IDs

Usage

annot(
  symbollist,
  species = NULL,
  ortholog = F,
  dboutput = "ncbi",
  idtype = NULL
)

Arguments

symbollist	character list of IDs or Symbols
species	character for species hs mm rn dr
ortholog	logical return homo sapiens ortholog of species
dboutput	character database used for Symbol annotation ncbi or ebi
idtype	character annotation database ID type among SYMBOL (by defaut) GENE, ENST, ENSG, ENSP, UNIPROT

Details

supported is : symbol, ncbi gene, ensembl gene , transcrit, protein, uniprot swissrot, uniprot trembl species : hs homo sapien , mm mus musculus , rn rattus norvegicus, dr danio rerio

Value

data.frame

Author(s)

Florent Dumont [email protected]

Examples

# not run
# annot(SymbolList)

---

Boxplot

Description

Boxplot

Usage

boxplot(
  dat,
  factor,
  outline = FALSE,
  title = "Boxplot",
  legendtitle = "TREATMENT",
  outlier = T,
  coefiqr = 1.5,
  ggplot = FALSE
)

Arguments

dat	matrix numeric
factor	factor
outline	logical display outliers FALSE by default
title	character
legendtitle	character
outlier	boolean
coefiqr	numeric
ggplot	logical use graphics library or ggplot FALSE by default

Details

To make boxplot from matrix.

Value

plot

Author(s)

Florent Dumont [email protected]

Examples

# not run
# mat1 %>% boxplot( factor = sif1$F3 )

---

boxplot for one var

Description

boxplot for one var

Usage

boxplot1(dat, ylab = "y", xlab = "TREATMENT", log = T)

Arguments

dat	data.frame
ylab	character
xlab	character
log	logical if TRUE data are delog in base 2

Value

plot

Author(s)

Florent Dumont [email protected]

Examples

# not run

---

MSigDB enrichment analysis

Description

MSigDB enrichment analysis

Usage

ena(
  SymbolList = NULL,
  geneannot = NULL,
  species = "hs",
  bg = 25000,
  filtergeneset = "all",
  overlapmin = 2,
  enaScoremin = 1,
  top = 80,
  labsize = 11,
  dpibarplot = "screen",
  path = ".",
  dirname = NULL
)

Arguments

SymbolList	character Symbol or NCBI gene ID
geneannot	data.frame
species	character hs mm rn dr ss
bg	numeric
filtergeneset	regexp to filter geneset database
overlapmin	numeric for minimum overlap between geneset and list
enaScoremin	numeric for minimum ratio ena
top	numeric top features to plot
labsize	numeric size of function in barplot
dpibarplot	character barplot resolution
path	character for relative path of output directory
dirname	character name for output

Value

file with enrichment analysis results

Author(s)

Florent Dumont [email protected]

Examples

# not run
# ena( Symbollist , filtergeneset = "reactome")

---

loading regular libraries

Description

load magrittr, dplyr, gplots, ggplot2, foreach, parallel, doParallel

Usage

env()

---

Hierarchical clustering classification

Description

make a hierarchical clustering classification

Usage

hc(
  dat,
  factor = NULL,
  title = "Hierarchical Clustering",
  plot = TRUE,
  method = "complete",
  legendtitle = "TREATMENT",
  cexlabel = 0.6
)

Arguments

dat	matrix numeric
factor	factor
title	character
plot	logical
method	character to choose agglomerative method of clustering dendrogramm
legendtitle	character
cexlabel	numeric

Details

Possible agglomerative method are the same hclust fonction : "complete" method by default

Value

a dendrogram

Author(s)

Florent Dumont [email protected]

Examples

# not run
# hc(dat)

---

Heatmap

Description

To make a heatmap

Usage

heatmap(
  dat,
  factor,
  method = "complete",
  dendrogram = "both",
  k = NULL,
  labCol = "",
  cexCol = 0.85,
  labRow = "",
  cexRow = 0.35,
  cexlegend = 0.65,
  keysize = 0.9,
  keycolor = c("darkgreen", "orange", "darkred"),
  parmar = c(5, 4, 5, 6)
)

Arguments

dat	matrix numeric
factor	factor
method	character
dendrogram	character to display 'none', 'row', 'column' or 'both' (by default) dendrograms
k	numeric number of clusters to colorize for rows
labCol	character
cexCol	numeric
labRow	Character
cexRow	numeric
cexlegend	numeric
keysize	numeric
keycolor	character of 3 for low mid high value of the key
parmar	numeric 4 values for margin sizes

Details

To make a heatmap from a matrix or a data.frame

Value

no returned value

Author(s)

Florent Dumont [email protected]

Examples

# not run
# library(magrittr)
# data(sif1)
# data(mat1)
# mat1 %>% heatmap(sif1$F3)

---

import tab file in data.frame

Description

import tab file in data.frame

Usage

input(filename, sep = "\t", quote = "")

Arguments

filename	character path to the file to read
sep	character for field separator
quote	character for field quote

Details

wrapper of read.table function for tabular separated files

Value

data.frame

Author(s)

Florent Dumont [email protected]

Examples

# not run
# input( "filename" ) -> dt

---

Normalization

Description

quantile normalization and log2

Usage

norm(dat, method = NULL, log = TRUE)

Arguments

dat	data.frame
method	character apply quantile normalization by default see details
log	logical apply log base 2

Details

for .method see limma normalizeBetweenArrays method

Value

data.frame

Author(s)

Florent Dumont [email protected]

Examples

# not run
# norm(dt)

---

Omic bioanalysis workflow

Description

Biostatistics analysis: QCs, ANOVA, threshold filtering, venn diagramm, cluster analysis for significant rows, pattern search

Graphics: volcanoplots, heatmaps, lineplots, boxplots (with Kruskal-Wallis test)

Functional analysis: MSigDB enrichment analysis, stringDB protein interaction network, topGO analysis (gene ontology)

See example section to test workflow with internal data.

Usage

omic(
  dat,
  sif,
  annot = NULL,
  species = "hs",
  model = NULL,
  paired = NULL,
  nested = NULL,
  batch = NULL,
  addfactor = NULL,
  qcs = TRUE,
  threshold = c(c(2, 6, 11), c(2, 6, 11) + 120),
  padj = "none",
  pattern = TRUE,
  venn = TRUE,
  cluster = TRUE,
  nc = c(2, 3, 6, 12),
  heatmap = TRUE,
  maxheatmap = NULL,
  volcanoplot = TRUE,
  lineplot = TRUE,
  boxplotrow = TRUE,
  ena = TRUE,
  enamin = 2,
  filtergeneset = "all",
  enaScoremin = 1.1,
  bg = 25000,
  sample = NULL,
  dopar = NULL,
  path = ".",
  dirname = NULL,
  zip = FALSE,
  remove = FALSE
)

Arguments

dat	data.frame normalize data table
sif	data.frame sample information file including model factors
annot	data.frame annotation with Symbol column for functional analysis
species	character available species: bt ce dr dm gg hs mm pt rn ss xt
model	character anova model factors (see details)
paired	character factor for paired design
nested	character factor for nested design
batch	character factor for batch effect design
addfactor	character additionnal factors
qcs	logical quality controls
threshold	numeric vector from 1 to 160 (see details)
padj	character fdr by defaut for Benjamini-Hochberg false discovery correction
pattern	logical search relevant pattern accross comparisons (see details)
venn	logical venn diagram
cluster	logical row hierarchical clustering using pearson correlation
nc	numeric number of clusters to cut in dendrogramm
heatmap	logical do heatmaps for all lists
maxheatmap	numeric max rows for heatmap
volcanoplot	logical make volcanoplot for each threshold
lineplot	logical do lineplot for significant features
boxplotrow	logical do boxplot for significant features with Kruskal
ena	logical msigdb enrichement analysis (over-representation analysis)
enamin	numeric min list size for functional analysis
filtergeneset	character regular expression to filter collection geneset (e.g. "reactome|tft")
enaScoremin	numeric for minimum ratio ena
bg	numeric background used for over-representation test
sample	numeric subset analysis
dopar	numeric core number
path	character results directory path
dirname	character results directory name
zip	logical compress results directory if TRUE
remove	logical remove uncompress results directory if TRUE

Details

Use uem915::env() to load required libraries before uem915::omic() (see example)

Accepted values for threshold param (1 to 150): see list -> uem915:::thresholdlist %>% lapply("[",c(1,2)) %>% unlist %>% matrix(ncol=2,byrow = T) %>% data.frame %>% setNames(c("pval","fc"))

Pattern: search relevant profiles among up and down comparison combinations

filtergeneset param: see geneset collections -> moalannotgene::genesetdb %>% lapply(names)

Experimental design examples:

• model = "TREAMTENT" for 1-way anova

• model = "TREATMENT+TIME+TREATMENT*TIME" for 3-ways anova with interaction

• model = "TREATMENT", paired = "CASE" for 2-ways paired anova

• model = "TREATMENT", batch = "BATCH" for 2-ways anova with remove batch effect (<=> paired anova)

• model = "TREATMENT+PHENOTYPE" for 2-ways anova

• model = "TREATMENT", addfactor = "PHENOTYPE" for 2-ways anova but venn, cluster and pattern are not applied to addfactor

• model = "TREATMENT", nested = "TREATMENTinCASE" for 2-ways nested design

• model = "TREATMENT", paired = "CASE", batch = "BATCH" for 3-ways paired anova with remove batch effect

Limitations:

• Complete block designs only

• Use dopar = 2 to decrease computing resources

• sample param will subsets random rows in dat and decrease analysis time.

Annotation updates: 05112023 for gene and ensembl, MSigDB 7.5.1, StringDB 12.0

Input format:

• Use uem915::input() to load external data from tsv files

• norm data table must contains IDs in first column

• norm data (in columns) and sample information (in rows) must have same order

• norm data and annotation rows must have same order

• use uem915::annot() function to annotate IDs (Symbols, ensembl and gene ids accepted)

Value

omic results directory

Author(s)

Florent Dumont [email protected]

Examples

# # test omic() with internal dataset GSE65055:
# # loading libraries
# library(uem915)
# uem915::env()
# # loading norm data
# moal:::GSE65055normdata -> normdata
# normdata %>% head
# # loading sample information file
# moal:::GSE65055sampledata -> sampledata
# sampledata %>% head
# # ordering factor levels
# sampledata$ANEUPLOIDY %>% ordered( c("Control","T13","T18","T21") ) -> sampledata$ANEUPLOIDY
# sampledata$TISSUE %>% as.factor -> sampledata$TISSUE
# # annotation
# normdata$rowID %>% moal::annot( species = "hs", idtype = "GENE" ) -> annotdata
# # omic analysis
# moal::omic(
#  dat = normdata, sif = sampledata, annot = annotdata, species = "hs",
#  model = "ANEUPLOIDY", batch = "TISSUE", threshold = c(6,126),
#  heatmap = T, lineplot = T, boxplotrow = T,
#  venn = F, cluster = F, pattern = F,
#  ena = T, network = F, topgo = F,
#  sample = NULL, dopar = NULL, zip = F,
#  dirname = "test", path = "." )

---

export data.frame in tab file

Description

export data.frame in tab file

Usage

output(dt, filename)

Arguments

dt	data.frame
filename	character

Author(s)

Florent Dumont [email protected]

Examples

# not run
# output( dt )

---

Quality Controls

Description

do descriptive statistics : histogram, boxplot , hierarchical clustering and ACP for column

Usage

qc(
  dat,
  sif = NULL,
  inputdata = F,
  histo = TRUE,
  boxplot = TRUE,
  hc = TRUE,
  acp = TRUE,
  dirname = NULL,
  path = "."
)

Arguments

dat	matrix numeric
sif	data.frame
inputdata	logical to export input data or not
histo	logical do histogram if TRUE by defaut
boxplot	logical do boxplot if TRUE by defaut
hc	logical do hierarchical clustering if TRUE by defaut
acp	logical do principal component analysis if TRUE by defaut
dirname	character
path	character

Details

return pval for each factor of anova model function use doparalle

Value

data.frame

Author(s)

Florent Dumont [email protected]

Examples

# not run
# qc(dat, metadata)

---

replace value by group median

Description

replace value by group median

Usage

replacegroupmed(dat, value = 0, factor)

Arguments

dat	character vector of mzxml file
value	numeric value to substitute
factor	character

Details

replace value by column median if group size is one and replace by row group median if group size > 1

Value

data.frame

Author(s)

Florent Dumont [email protected]

Examples

# not run
# replacegroupmed(dat)

---

Venn diagramm

Description

To make a Venn diagramm of 2, 3 or 4 lists

Usage

venn(
  list = NULL,
  listnames = NULL,
  returnlist = F,
  title = "Venn Diagram",
  plot = T,
  export = F,
  path = ".",
  dirname = "venn"
)

Arguments

list	list of 2 , 3 or 4 character vector or list of two data.frame to compare
listnames	character list names to display on graph
returnlist	logical
title	character title to display on graph
plot	logical to display the plot or not
export	logical export list in file
path	character
dirname	character name of the directory created when export = T

Details

until 4 list

Value

venn plot and new lists generated by venn.

Author(s)

Florent Dumont [email protected]

Examples

# not run
# library(magrittr)
# list(
#   c(letters[6:20] , letters[25] ) ,
#   letters[1:15] ,
#   c( letters[2:5] , letters[8:23] ) ) %>% venn

---