%\VignetteIndexEntry{NeighborNet}
%\VignetteDepends{NeighborNet}
%\VignettePackage{NeighborNet}


\documentclass[11pt]{article}

\usepackage{hyperref}
\usepackage[margin=1in]{geometry}

\newcommand{\Rfunction}[1]{{\texttt{#1}}}
\newcommand{\Rmethod}[1]{{\texttt{#1}}}

\newcommand{\Robject}[1]{{\texttt{#1}}}
\newcommand{\Rpackage}[1]{{\textit{#1}}}
\newcommand{\Rclass}[1]{{\textit{#1}}}


\begin{document}
\SweaveOpts{concordance=TRUE}
\title{Neighbor\_net analysis}
\author{Sahar Ansari and Sorin Draghici\\
Department of Computer Science, Wayne State University, Detroit MI 48201}
\maketitle

\begin{abstract}

This package is indented to be the R implementation of the method introduced in \cite{ansari2017approach}. Neighbor\_net analysis aims to take advantage of the prior knowledge of gene-gene interactions and idetifies the putative mechanisms at play in the given condition (e.g. a disease, a treatment, etc.). The captured network can be useful for the prediction of mechanisms of action of drugs or the responses of an organism to a specific impact.

\end{abstract}

\section{Neighbor\_net analysis}

Neighbor\_net (\Rmethod{neighborNet}) is a tool to identify the active mechanism involved in an investigated phenotype.
This method uses two sources of data: one is the experiment data and the other is the gene-gene interactions knowledge.

\subsection{Gene-gene interaction knowledge}

Neighbor\_net can accept any gene-gene interaction information obtained from different databases.
The data has to be converted to a list format. Each element in the list represents the neighborhood of one gene.

We provided an example that includes the interactions exist in KEGG\cite{Kanehisa:2000} and HPRD~\cite{Peri:2003} databases.

<<eval=TRUE, echo=TRUE>>=
load(system.file("extdata/listofgenes.RData", package = "NeighborNet"))
@

The \Robject{listofgenes} is a list including the neighbors of the genes in the analysis:
<<eval=TRUE, echo=TRUE>>=
head(listofgenes)
@




\subsection{Experiment data}

As an example, we provided five pre-processed data sets from GEO (GSE4183, GSE9348, GSE21510, GSE32323, GSEl8671).

These data study the expression change between colorectal cancer and normal patients. The data was preprocessed using the \Rpackage{limma} package. Only probe sets with a gene associated to them have been kept.

<<eval=TRUE, echo=TRUE>>=
load(system.file("extdata/dataColorectal4183.RData", package = "NeighborNet"))
load(system.file("extdata/dataColorectal9348.RData", package = "NeighborNet"))
load(system.file("extdata/dataColorectal21510.RData", package = "NeighborNet"))
load(system.file("extdata/dataColorectal32323.RData", package = "NeighborNet"))
load(system.file("extdata/dataColorectal8671.RData", package = "NeighborNet"))
head(dataColorectal4183)
@


The next step is to select the genes that are differentially expressed, with p-value lower than 1\% and absolute fold change more than 1.5.

<<eval=TRUE, echo=TRUE>>=
pvThreshold <- 0.01
foldThreshold <- 1.5
de1 <- dataColorectal4183$EntrezID [
  dataColorectal4183$adj.P.Val < pvThreshold &
  abs(dataColorectal4183$logFC) > foldThreshold
]
de2 <- dataColorectal9348$EntrezID [
  dataColorectal9348$adj.P.Val < pvThreshold &
  abs(dataColorectal9348$logFC) > foldThreshold
]
de3 <- dataColorectal21510$EntrezID [
  dataColorectal21510$adj.P.Val < pvThreshold &
  abs(dataColorectal21510$logFC) > foldThreshold
]
de4 <- dataColorectal32323$EntrezID [
  dataColorectal32323$adj.P.Val < pvThreshold &
  abs(dataColorectal32323$logFC) > foldThreshold
]
de5 <- dataColorectal8671$EntrezID [
  dataColorectal8671$adj.P.Val < pvThreshold &
  abs(dataColorectal8671$logFC) > foldThreshold
]

@

Later, the differentialy expressed genes from different datasets should be combined together:
<<eval=TRUE, echo=TRUE>>=
de <- unique( c(de1,de2,de3,de4,de5))
@

The reference contains all the genes measured in the analysis:
<<eval=TRUE, echo=TRUE>>=
ref <- unique( c(
  dataColorectal4183$EntrezID,
  dataColorectal9348$EntrezID,
  dataColorectal21510$EntrezID,
  dataColorectal32323$EntrezID,
  dataColorectal8671$EntrezID
))
head(ref)
@

\subsection{Neighbor\_net analysis and resulted network}

We have all the input fot Neighbor\_net analysis.

\begin{itemize}
\item the gene-gene knowledge in a list format -\Robject {listofgenes}
\item the experiment data -\Robject{de} and -\Robject{ref}
\end{itemize}


<<eval=TRUE, echo=TRUE>>=
library("NeighborNet")
library("graph")
sig_genes <- neighborNet(de = de, ref = ref, listofgenes=listofgenes)
sig_genes
@

\subsection{Graphical representation of results}

To visualize the identified network use the function \Rfunction{plot}(see Fig.~\ref{fig:colorectalpathway}):


<<label=peRes_twoway1,include=FALSE>>=
require("graph")
attrs <- list(node= list(fontsize=40,fixedsize= FALSE),graph=list(overlap=FALSE), edge=list(lwd=0.6))
nAttr <- list()
nAttr$color <- c(rep("white",length(nodes(sig_genes))))
names(nAttr$color) <- nodes(sig_genes)
plot(sig_genes)
@

\begin{figure}
\begin{center}
\resizebox{!}{4in}{
\includegraphics{fig1}
}
\end{center}
\caption{The active network that describes the putative mechanism involved in colorectal cancer.}
\label{fig:colorectalpathway}
\end{figure}



\bibliographystyle{abbrv}
\bibliography{neighborNet}

\end{document}