library(Linnorm)
Normalized <- Linnorm.Norm(Islam2011)

# Important: depending on downstream analysis, please
# set output to be "XPM" or "Raw".
# Set to "XPM" if downstream tools will convert the 
# dataset into CPM or TPM.
# Set to "Raw" if input is raw counts and downstream 
# tools will work with raw counts.

# You can use this file with Excel.
write.table(Normalized, "Normalized.txt",
quote=FALSE, sep="\t", col.names=TRUE, row.names=TRUE)

library(Linnorm)
data(LIHC)
datamatrix <- LIHC

# 5 samples for condition 1 and 5 samples for condition 2.
# You might need to edit this line
design <- c(rep(1,5),rep(2,5))
# These lines can be copied directly.
design <- model.matrix(~ 0+factor(design))
colnames(design) <- c("group1", "group2")
rownames(design) <- colnames(datamatrix)

library(Linnorm)
# The Linnorm-limma pipeline only consists of one line.
DEG_Results <- Linnorm.limma(datamatrix,design)

# The DEG_Results matrix contains your DEG analysis results.

library(Linnorm)
# Just add output="Both" into the argument list
BothResults <- Linnorm.limma(datamatrix,design,output="Both")

# To separate results into two matrices:
DEG_Results <- BothResults$DEResults
TransformedMatrix <- BothResults$Linnorm
# The DEG_Results matrix now contains DEG analysis results.
# The TransformedMatrix matrix now contains a Linnorm
# Transformed dataset.

write.table(DEG_Results, "DEG_Results.txt", quote=FALSE, sep="\t",
col.names=TRUE,row.names=TRUE)

Genes10 <- DEG_Results[order(DEG_Results[,"adj.P.Val"]),][1:10,]
# Users can print the gene list by the following command:
# print(Genes10)

# logFC: log 2 fold change of the gene.
# XPM: If input is raw count or CPM, XPM is CPM.
#   If input is RPKM, FPKM or TPM, XPM is TPM.
# t: moderated t statistic.
# P.Value: p value.
# adj.P.Val: Adjusted p value. This is also called False Discovery Rate or q value.}
# B: log odds that the feature is differential.

# Note all columns are printed here

TheCluster <- which(results$Cluster[,1] == "Mmp2")

# Index of the genes
ListOfGenes <- which(results$Cluster[,2] == TheCluster)

# Names of the genes
GeneNames <- results$Cluster[ListOfGenes,1]

# Users can write these genes into a file for further analysis.

top100 <- results$Results[order(results$Results[,4],decreasing=FALSE)[1:100],1]

Top100.Cor.Matrix <- results$Cor.Matrix[top100,top100]

library(RColorBrewer)
library(gplots)

heatmap.2(as.matrix(Top100.Cor.Matrix),
# Hierarchical clustering on both row and column
Rowv=TRUE, Colv=TRUE,
# Center white color at correlation 0
symbreaks=TRUE,
# Turn off level trace
trace="none",
# x and y axis labels
xlab = 'Genes', ylab = "Genes",
# Turn off density info
density.info='none',
# Control color
key.ylab=NA,
col=colorRampPalette(c("blue", "white", "yellow"))(n = 1000),
lmat=rbind(c(4, 3), c(2, 1)),
# Gene name font size
cexRow=0.3,
cexCol=0.3,
# Margins
margins = c(8, 8)
)

tSNE.results <- Linnorm.tSNE(Islam2011[,1:92])

# Here, we can see two clusters.
print(tSNE.results$plot$plot)

# The first 48 samples belong to mouse embryonic stem cells.
Groups <- rep("ES_Cells",48)
# The next 44 samples are mouse embryonic fibroblasts.
Groups <- c(Groups, rep("EF_Cells",44))

# Useful arguments:
# Group:
# allows user to provide each sample's information.
# num_center:
# how many clusters are supposed to be there.
# num_PC:
# how many principal components should be used in k-means
# clustering.

tSNE.results <- Linnorm.tSNE(Islam2011[,1:92],
 Group=Groups, num_center=2)

# Here, we can see two clusters.
print(tSNE.results$plot$plot)

PCA.results <- Linnorm.PCA(Islam2011[,1:92])

# Here, we can see multiple clusters.
print(PCA.results$plot$plot)

# The first 48 samples belong to mouse embryonic stem cells.
Groups <- rep("ES_Cells",48)
# The next 44 samples are mouse embryonic fibroblasts.
Groups <- c(Groups, rep("EF_Cells",44))

# Useful arguments:
# Group:
# allows user to provide each sample's information.
# num_center:
# how many clusters are supposed to be there.
# num_PC
# how many principal components should be used in k-means
# clustering.

PCA.results <- Linnorm.PCA(Islam2011[,1:92],
Group=Groups, num_center=2, num_PC=3)

# Here, we can see two clusters.
print(PCA.results$plot$plot)

data(Islam2011)
Islam <- Islam2011[,1:92]

# 48 ESC, 44 EF, and 4 NegCtrl
Group <- c(rep("ESC",48),rep("EF",44))
colnames(Islam) <- paste(colnames(Islam),Group,sep="_")

# Note that there are 3 known clusters.
HClust.Results <- Linnorm.HClust(Islam,Group=Group,
num_Clust=2, fontsize=1.5, Color = c("Red","Blue"), RectColor="Green")

print(HClust.Results$plot$plot)

library(Linnorm)
data(SEQC)
SampleA <- SEQC

# This will generate two sets of RNA-seq data with 5 replicates each. 
# It will have 20000 genes totally with 2000 genes being differentially 
# expressed. It has the Negative Binomial distribution.
SimulatedData <- RnaXSim(SampleA)

# Index of differentially expressed genes.
DE_Index <- SimulatedData[[2]]

# Expression Matrix
ExpMatrix <- SimulatedData[[1]]

# Sample Set 1
Sample1 <- ExpMatrix[,1:3]

# Sample Set 2
Sample2 <- ExpMatrix[,4:6]

data(SEQC)
SampleA <- SEQC

library(Linnorm)
SimulatedData <- RnaXSim(SampleA,
distribution="Gamma", # Distribution in the simulated dataset.
# Put "NB" for Negative Binomial, "Gamma" for Gamma,
# "Poisson" for Poisson and "LogNorm" for Log Normal distribution.
NumRep=5, # Number of replicates in each sample set. 
# 5 will generate 10 samples in total.
NumDiff = 1000, # Number of differentially expressed genes.
NumFea = 5000 # Total number of genes in the dataset
)

# Index of differentially expressed genes.
DE_Index <- SimulatedData[[2]]

# Expression Matrix
ExpMatrix <- SimulatedData[[1]]

# Simulated Sample Set 1
Sample1 <- ExpMatrix[,1:3]

# Simulated Sample Set 2
Sample2 <- ExpMatrix[,4:6]