arvinUserGuide.tex

\documentclass[]{article}
\usepackage{lmodern}
\usepackage{amssymb,amsmath}
\usepackage{ifxetex,ifluatex}
\usepackage{fixltx2e} % provides \textsubscript
\ifnum 0\ifxetex 1\fi\ifluatex 1\fi=0 % if pdftex
  \usepackage[T1]{fontenc}
  \usepackage[utf8]{inputenc}
\else % if luatex or xelatex
  \ifxetex
    \usepackage{mathspec}
  \else
    \usepackage{fontspec}
  \fi
  \defaultfontfeatures{Ligatures=TeX,Scale=MatchLowercase}
\fi
% use upquote if available, for straight quotes in verbatim environments
\IfFileExists{upquote.sty}{\usepackage{upquote}}{}
% use microtype if available
\IfFileExists{microtype.sty}{%
\usepackage{microtype}
\UseMicrotypeSet[protrusion]{basicmath} % disable protrusion for tt fonts
}{}
\usepackage[margin=1in]{geometry}
\usepackage{hyperref}
\hypersetup{unicode=true,
            pdftitle={ARVIN : Identifying Risk Noncoding Variants Using Disease-relevant Gene Regulatory Networks},
            pdfauthor={Long Gao, Yasin Uzun, Kai Tan},
            pdfborder={0 0 0},
            breaklinks=true}
\urlstyle{same}  % don't use monospace font for urls
\usepackage{color}
\usepackage{fancyvrb}
\newcommand{\VerbBar}{|}
\newcommand{\VERB}{\Verb[commandchars=\\\{\}]}
\DefineVerbatimEnvironment{Highlighting}{Verbatim}{commandchars=\\\{\}}
% Add ',fontsize=\small' for more characters per line
\usepackage{framed}
\definecolor{shadecolor}{RGB}{248,248,248}
\newenvironment{Shaded}{\begin{snugshade}}{\end{snugshade}}
\newcommand{\KeywordTok}[1]{\textcolor[rgb]{0.13,0.29,0.53}{\textbf{#1}}}
\newcommand{\DataTypeTok}[1]{\textcolor[rgb]{0.13,0.29,0.53}{#1}}
\newcommand{\DecValTok}[1]{\textcolor[rgb]{0.00,0.00,0.81}{#1}}
\newcommand{\BaseNTok}[1]{\textcolor[rgb]{0.00,0.00,0.81}{#1}}
\newcommand{\FloatTok}[1]{\textcolor[rgb]{0.00,0.00,0.81}{#1}}
\newcommand{\ConstantTok}[1]{\textcolor[rgb]{0.00,0.00,0.00}{#1}}
\newcommand{\CharTok}[1]{\textcolor[rgb]{0.31,0.60,0.02}{#1}}
\newcommand{\SpecialCharTok}[1]{\textcolor[rgb]{0.00,0.00,0.00}{#1}}
\newcommand{\StringTok}[1]{\textcolor[rgb]{0.31,0.60,0.02}{#1}}
\newcommand{\VerbatimStringTok}[1]{\textcolor[rgb]{0.31,0.60,0.02}{#1}}
\newcommand{\SpecialStringTok}[1]{\textcolor[rgb]{0.31,0.60,0.02}{#1}}
\newcommand{\ImportTok}[1]{#1}
\newcommand{\CommentTok}[1]{\textcolor[rgb]{0.56,0.35,0.01}{\textit{#1}}}
\newcommand{\DocumentationTok}[1]{\textcolor[rgb]{0.56,0.35,0.01}{\textbf{\textit{#1}}}}
\newcommand{\AnnotationTok}[1]{\textcolor[rgb]{0.56,0.35,0.01}{\textbf{\textit{#1}}}}
\newcommand{\CommentVarTok}[1]{\textcolor[rgb]{0.56,0.35,0.01}{\textbf{\textit{#1}}}}
\newcommand{\OtherTok}[1]{\textcolor[rgb]{0.56,0.35,0.01}{#1}}
\newcommand{\FunctionTok}[1]{\textcolor[rgb]{0.00,0.00,0.00}{#1}}
\newcommand{\VariableTok}[1]{\textcolor[rgb]{0.00,0.00,0.00}{#1}}
\newcommand{\ControlFlowTok}[1]{\textcolor[rgb]{0.13,0.29,0.53}{\textbf{#1}}}
\newcommand{\OperatorTok}[1]{\textcolor[rgb]{0.81,0.36,0.00}{\textbf{#1}}}
\newcommand{\BuiltInTok}[1]{#1}
\newcommand{\ExtensionTok}[1]{#1}
\newcommand{\PreprocessorTok}[1]{\textcolor[rgb]{0.56,0.35,0.01}{\textit{#1}}}
\newcommand{\AttributeTok}[1]{\textcolor[rgb]{0.77,0.63,0.00}{#1}}
\newcommand{\RegionMarkerTok}[1]{#1}
\newcommand{\InformationTok}[1]{\textcolor[rgb]{0.56,0.35,0.01}{\textbf{\textit{#1}}}}
\newcommand{\WarningTok}[1]{\textcolor[rgb]{0.56,0.35,0.01}{\textbf{\textit{#1}}}}
\newcommand{\AlertTok}[1]{\textcolor[rgb]{0.94,0.16,0.16}{#1}}
\newcommand{\ErrorTok}[1]{\textcolor[rgb]{0.64,0.00,0.00}{\textbf{#1}}}
\newcommand{\NormalTok}[1]{#1}
\usepackage{graphicx,grffile}
\makeatletter
\def\maxwidth{\ifdim\Gin@nat@width>\linewidth\linewidth\else\Gin@nat@width\fi}
\def\maxheight{\ifdim\Gin@nat@height>\textheight\textheight\else\Gin@nat@height\fi}
\makeatother
% Scale images if necessary, so that they will not overflow the page
% margins by default, and it is still possible to overwrite the defaults
% using explicit options in \includegraphics[width, height, ...]{}
\setkeys{Gin}{width=\maxwidth,height=\maxheight,keepaspectratio}
\IfFileExists{parskip.sty}{%
\usepackage{parskip}
}{% else
\setlength{\parindent}{0pt}
\setlength{\parskip}{6pt plus 2pt minus 1pt}
}
\setlength{\emergencystretch}{3em}  % prevent overfull lines
\providecommand{\tightlist}{%
  \setlength{\itemsep}{0pt}\setlength{\parskip}{0pt}}
\setcounter{secnumdepth}{0}
% Redefines (sub)paragraphs to behave more like sections
\ifx\paragraph\undefined\else
\let\oldparagraph\paragraph
\renewcommand{\paragraph}[1]{\oldparagraph{#1}\mbox{}}
\fi
\ifx\subparagraph\undefined\else
\let\oldsubparagraph\subparagraph
\renewcommand{\subparagraph}[1]{\oldsubparagraph{#1}\mbox{}}
\fi

%%% Use protect on footnotes to avoid problems with footnotes in titles
\let\rmarkdownfootnote\footnote%
\def\footnote{\protect\rmarkdownfootnote}

%%% Change title format to be more compact
\usepackage{titling}

% Create subtitle command for use in maketitle
\newcommand{\subtitle}[1]{
  \posttitle{
    \begin{center}\large#1\end{center}
    }
}

\setlength{\droptitle}{-2em}
  \title{ARVIN : Identifying Risk Noncoding Variants Using Disease-relevant Gene
Regulatory Networks}
  \pretitle{\vspace{\droptitle}\centering\huge}
  \posttitle{\par}
  \author{Long Gao, Yasin Uzun, Kai Tan}
  \preauthor{\centering\large\emph}
  \postauthor{\par}
  \predate{\centering\large\emph}
  \postdate{\par}
  \date{May 18, 2018}


\begin{document}
\maketitle

{
\setcounter{tocdepth}{4}
\tableofcontents
}
\subsection{1 Introduction}\label{introduction}

Identifying causal noncoding variants remains a daunting task. Because
noncoding variants exert their effects in the context of a gene
regulatory network (GRN), we hypothesize that explicit use of
disease-relevant GRN can significantly improve the inference accuracy of
noncoding risk variants. We describe Annotation of Regulatory Variants
using Integrated Networks (ARVIN), a general computational framework for
predicting causal noncoding variants. For each disease, ARVIN first
constructs a GRN using multi-dimensional omics data oncell/tissue-type
relevant to the disease. ARVIN then uses a set of novel regulatory
network-based features, combined with sequence-based features to make
predictions.

This user guide explains ARVIN package. If you want to make a quick
start and run ARVIN, please refer to the related document in this
repository (ARVIN\_Quick\_Start.docx).

ARVIN uses genome annotation and motif data. You need to download
\url{http://tanlab4generegulation.org/arvin_annotation_data.tar.gz} to
run ARVIN. In addition, please make sure following dependency packages
also installed including ``caret'', ``randomForest'', and ``igraph''.

\subsection{2 Network construction}\label{network-construction}

As a first step of ARVIN, an integrative GRN for each disease-relevant
cell/tissue type should be constructed. In this network, there are two
types of nodes which are genes and snps/variants as well as two types of
edges which are gene-gene interacstions and snp-gene interactions. Gene
interaction network is used as the network backbone, and then this
backbone is integrated with enhancer-promoter interactions(snp-gene
interactions). For each type of nodes and edges, normalized scores are
computed using cell/tissue speicific information.

\subsubsection{2.1 Enhancer-promoter
interaction}\label{enhancer-promoter-interaction}

ARVIN uses enhancer-promoter interaction for mapping SNPS to genes via
enhancers. The enhancer-promoter interaction data must be in tab
separated format as follows:\\
\#Chr Start End Target Score\\
chr9 22124001 22126001 ENST00000452276 0.93\\
chr2 242792001 242794001 ENST00000485966 0.792\\
You can use IM-PET software for predicting enhancer-promoter
interactions. For this purpose, you can download IM-PET from
\url{http://tanlab4generegulation.org/IM-PET.html} . IM-PET uses
enhancer predictions (computed using CSI-ANN) and gene expression data
to compute enhancer-promoter interactions. You can use the output of
IM-PET (which is as shown above) as input for ARVIN.

\subsubsection{2.2 Obtain gene-gene interaction
network}\label{obtain-gene-gene-interaction-network}

Gene interaction network can be obtained from multiple sources including
protein-protein interaction networks and functional interaction networks
such as BioGRID, BIND, HumantNet, STRING and etc.

\subsubsection{2.3 Network scoring}\label{network-scoring}

To make different types of scores comparable, we used a min-max
normalization to normalize scores within each category. For gene
interaction network, the interaction score indicating how strong the
interactions between genes can be directly used from corresponding
database followed by normalization. SNP-gene interaction score can be
assigned with enhancer-promoter score or scores indicating the
interaction strength between enhancer and target promoters. SNPs weight
can be represented by TF binding disruption score computed using our
script/function. Genes can be weighted using differential expression
information.

\subsubsection{2.4 Network input file
format}\label{network-input-file-format}

There are two types of network input files users need to prepare. One is
the node attribute file and the other is network/edge attribute file.
The node attribute file has 3 columns. The first column denotes snp id
or gene id, and the second column denotes the score of this snp or gene.
The third column specifiy if this node is a snp or gene. In the network
file, there are 4 columns. The first two columns list two nodes of a
given edge. The third collumn has the normalized score for this edge.
The forth column specifies edge type indicating if this interaction is
between snps and genes or genes.

\begin{Shaded}
\begin{Highlighting}[]
\NormalTok{edgeFile <-}\StringTok{ "example_1/EdgeFile.txt"}
\NormalTok{nodeFile <-}\StringTok{ "example_1/NodeFile.txt"}
\NormalTok{SNP_pFile <-}\StringTok{ "example_1/snp_pval.txt"}
\NormalTok{edge_data <-}\StringTok{ }\KeywordTok{read.table}\NormalTok{(edgeFile, }\DataTypeTok{sep=}\StringTok{"}\CharTok{\textbackslash{}t}\StringTok{"}\NormalTok{)}
\NormalTok{node_data   <-}\StringTok{ }\KeywordTok{read.table}\NormalTok{(nodeFile, }\DataTypeTok{sep=}\StringTok{"}\CharTok{\textbackslash{}t}\StringTok{"}\NormalTok{)}
\KeywordTok{colnames}\NormalTok{(node_data) <-}\StringTok{ }\KeywordTok{c}\NormalTok{(}\StringTok{"Node"}\NormalTok{, }\StringTok{"Node score"}\NormalTok{, }\StringTok{"Node type"}\NormalTok{)}
\KeywordTok{colnames}\NormalTok{(edge_data) <-}\StringTok{ }\KeywordTok{c}\NormalTok{(}\StringTok{"First node"}\NormalTok{, }\StringTok{"Second node"}\NormalTok{, }\StringTok{"Edge score"}\NormalTok{, }\StringTok{"Edge type"}\NormalTok{)}
\NormalTok{edge_data[}\DecValTok{98}\OperatorTok{:}\DecValTok{103}\NormalTok{,]}
\end{Highlighting}
\end{Shaded}

\begin{verbatim}
##     First node Second node Edge score Edge type
## 98    CR004576        3043  0.7854000        EP
## 99    CR034843        3043  0.3298000        EP
## 100   CR040152        3043  0.6225000        EP
## 101       2237        5111  0.9965896        FI
## 102        506         509  0.9948976        FI
## 103        498         506  0.9827500        FI
\end{verbatim}

\begin{Shaded}
\begin{Highlighting}[]
\NormalTok{node_data[}\DecValTok{98}\OperatorTok{:}\DecValTok{103}\NormalTok{,]}
\end{Highlighting}
\end{Shaded}

\begin{verbatim}
##         Node Node score Node type
## 98  CR004576 0.79200000      eSNP
## 99  CR034843 0.80740000      eSNP
## 100 CR040152 0.76540000      eSNP
## 101       19 0.70045366      Gene
## 102       20 0.14164698      Gene
## 103       24 0.07013084      Gene
\end{verbatim}

\subsection{3 Prepare features for risk variants
prediction}\label{prepare-features-for-risk-variants-prediction}

\subsubsection{3.1 Network-based features}\label{network-based-features}

For most of network features such as the centrality, we wrapped up
functions from ``igraph'' package to calculate their values. We also
implemented our module identification algorithm to find modules
containing snps we are intereted in. To calculate all network based
features, users can simply call NetFeature(). SNPs in the same enhancer
usually have similar topological features. However, we can further
distinguish them using their TF motif breaking scores.

\begin{Shaded}
\begin{Highlighting}[]
\NormalTok{Nodes <-}\StringTok{ }\KeywordTok{as.character}\NormalTok{(edge_data[,}\DecValTok{2}\NormalTok{])}
\NormalTok{Net <-}\StringTok{ }\KeywordTok{makeNet}\NormalTok{(edgeFile, nodeFile)}
\NormalTok{topoFeature <-}\KeywordTok{NetFeature}\NormalTok{(Net, nodeFile, edgeFile, SNP_pFile)}
\KeywordTok{head}\NormalTok{(topoFeature)}
\end{Highlighting}
\end{Shaded}

\begin{verbatim}
##           Module_Score Betweenness Closeness  Pagerank Weighted_Degree
## CR080767      2.602336   20.894963 0.2186483 11.166083       24.734249
## CR083996      1.842770    4.278196 0.2139707  3.212460        6.388578
## CR0911347     2.043153   22.710396 0.2209625  5.589152        8.799405
## CR0911356     2.300549    1.998601 0.2116995  2.038007        4.594794
## CR095246      2.202802   48.855359 0.2212212 31.829972       85.214685
## CR095443      1.597560    8.052195 0.2143816  6.083444       11.806411
##           SNP_Disrutption
## CR080767       0.97255655
## CR083996       0.53409931
## CR0911347      0.81091625
## CR0911356      0.65090651
## CR095246       0.77098111
## CR095443       0.03455832
\end{verbatim}

\paragraph{3.1.1 Betweenness centrality}\label{betweenness-centrality}

Betweenness is a centrality measure of a vertex within a graph.
Betweenness centrality quantifies the number of times a node acts as a
bridge along the shortest path between two other nodes.

\begin{Shaded}
\begin{Highlighting}[]
\NormalTok{bet_vals <-}\StringTok{ }\KeywordTok{BetFeature}\NormalTok{(Net, edge_data)}
\KeywordTok{head}\NormalTok{(bet_vals)}
\end{Highlighting}
\end{Shaded}

\begin{verbatim}
##  CR080767  CR083996 CR0911347 CR0911356  CR095246  CR095443 
## 20.894963  4.278196 22.710396  1.998601 48.855359  8.052195
\end{verbatim}

\paragraph{3.1.2 Closeness centrality}\label{closeness-centrality}

Closeness is a measure of the degree to which an individual is near all
other individuals in a network. It is the inverse of the sum of the
shortest distances between each node and every other node in the
network. Closeness is the reciprocal of farness.

\begin{Shaded}
\begin{Highlighting}[]
\NormalTok{close_vals <-}\StringTok{ }\KeywordTok{CloseFeature}\NormalTok{(Net, edge_data)}
\KeywordTok{head}\NormalTok{(close_vals)}
\end{Highlighting}
\end{Shaded}

\begin{verbatim}
##  CR080767  CR083996 CR0911347 CR0911356  CR095246  CR095443 
## 0.2186483 0.2139707 0.2209625 0.2116995 0.2212212 0.2143816
\end{verbatim}

\paragraph{3.1.3 Pagerank centrality}\label{pagerank-centrality}

PageRank (PR) is an algorithm used by Google Search to rank websites in
their search engine results. PageRank is a way of measuring the
importance of website pages. In the biological networks, we can also use
this algorithm to measure the importance of genes/nodes.

\begin{Shaded}
\begin{Highlighting}[]
\NormalTok{page_vals <-}\StringTok{ }\KeywordTok{PageFeature}\NormalTok{(Net, edge_data)}
\KeywordTok{head}\NormalTok{(page_vals)}
\end{Highlighting}
\end{Shaded}

\begin{verbatim}
##  CR080767  CR083996 CR0911347 CR0911356  CR095246  CR095443 
## 11.166083  3.212460  5.589152  2.038007 31.829972  6.083444
\end{verbatim}

\paragraph{3.1.4 Weighted degree}\label{weighted-degree}

The weighted degree of a node is like the degree. It's based on the
number of edge for a node, but ponderated by the weigtht of each edge.
It's doing the sum of the weight of the edges.

\begin{Shaded}
\begin{Highlighting}[]
\NormalTok{wd_vals <-}\StringTok{ }\KeywordTok{WDFeature}\NormalTok{(Adj_List, edge_data)}
\KeywordTok{head}\NormalTok{(wd_vals)}
\end{Highlighting}
\end{Shaded}

\begin{verbatim}
##  CR080767  CR083996 CR0911347 CR0911356  CR095246  CR095443 
## 24.734249  6.388578  8.799405  4.594794 85.214685 11.806411
\end{verbatim}

\paragraph{3.1.5 Module score}\label{module-score}

Gene modules downstream of an eSNP. Our overall hypothesis is that a
causal eSNP contributes to disease risk by directly causing expression
changes in genes of diseaserelevant pathways. Thus, in addition to the
direct target gene of the eSNP, other genes in the same pathway can also
provide discriminative information. With the weighted GRN, our goal is
to identify ``heavy'' gene modules in the network that connects a given
eSNP to a set of genes

\begin{Shaded}
\begin{Highlighting}[]
\NormalTok{mod_vals <-}\StringTok{ }\KeywordTok{ModuleFeature}\NormalTok{(Adj_List, E_adj, eSNP_seeds, V_weight, Nodes)}
\KeywordTok{head}\NormalTok{(mod_vals)}
\end{Highlighting}
\end{Shaded}

\begin{verbatim}
##  CR080767  CR083996 CR0911347 CR0911356  CR095246  CR095443 
##  2.602336  1.842770  2.043153  2.300549  2.202802  1.597560
\end{verbatim}

\subsubsection{3.2 GWAVA features}\label{gwava-features}

ARVIN uses sequence features for the input SNPs generated by GWAVA.
GWAVA is an open-source software developed by Sanger Institute. You can
either upload the SNPs to GWAVA web page and get the output or download
the source and run locally. For running GWAVA online navigate to
\url{https://www.sanger.ac.uk/sanger/StatGen_Gwava}, upload the list of
input SNPs and get the features in csv format, which will be input for
ARVIN. If you prefer to run it locally, you need to dowload the source
code from
\url{ftp://ftp.sanger.ac.uk/pub/resources/software/gwava/v1.0/src/} and
annotation data from
\url{ftp://ftp.sanger.ac.uk/pub/resources/software/gwava/v1.0/source_data/}
. Then you can run it local by running gwava\_annotate.py and generate
the features, which will be input for ARVIN.

\begin{Shaded}
\begin{Highlighting}[]
\NormalTok{gwava <-}\StringTok{ }\KeywordTok{read.table}\NormalTok{(}\StringTok{"example_1/gwava_matrix.txt"}\NormalTok{, }\DataTypeTok{header=}\NormalTok{T, }\DataTypeTok{sep=}\StringTok{"}\CharTok{\textbackslash{}t}\StringTok{"}\NormalTok{)}
\NormalTok{gwava[}\DecValTok{1}\OperatorTok{:}\DecValTok{8}\NormalTok{,}\DecValTok{168}\OperatorTok{:}\DecValTok{174}\NormalTok{]}
\end{Highlighting}
\end{Shaded}

\begin{verbatim}
##             avg_daf in_cpg seq_A seq_C seq_G seq_T repeat.
## CR080767  0.0015376      1     0     1     0     0       0
## CR083996  0.0016738      0     0     1     0     0       0
## CR0911347 0.0021400      0     0     0     1     0       0
## CR0911356 0.0013821      0     0     0     0     1       1
## CR095246  0.0006147      0     0     1     0     0       0
## CR095443  0.0006415      0     0     0     1     0       0
## CR109943  0.0019969      0     0     0     0     1       1
## CR1210014 0.0017969      0     1     0     0     0       0
\end{verbatim}

\subsubsection{3.3 FunSeq features}\label{funseq-features}

ARVIN also uses sequence features generated by FunSeq. FunSeq can also
be run online or binaries can be downloaded to run locally. For running
FunSeq online, navigate to \url{http://funseq.gersteinlab.org/analysis}
and upload the list of SNPs that you want to analyze. In the web page,
it is noted that the input SNPs can be uploaded in bed format, SNP
coordinates followed by reference and alternate alleles; but we
discovered that it fails to process bed input. In order to have it run,
you the first two seperators need to be two spaces and last two
separators need to be tabs, as follows:\\
chr16··4526757··4526758 G A\\
chr14··52733136··52733137 C A\\
where each dot (·) represents a space. Then, FunSeq will generate the
features by selecting ``bed'' as the output format., which will be used
as input by ARVIN.

If you prefer to run FunSeq locally, you can download FunSeq binaries
from \url{http://funseq.gersteinlab.org/static/funseq-0.1.tar.gz} and
extract it into your local. You will also need to download FunSeq
annotation data from
\url{http://funseq.gersteinlab.org/static/data/data.tar.gz} , extract it
into directory that you saved the binaries. Then you can run FunSeq
binary file by setting the output format to bed.

\begin{Shaded}
\begin{Highlighting}[]
\NormalTok{funseq <-}\StringTok{ }\KeywordTok{read.table}\NormalTok{(}\StringTok{"example_1/funseq_matrix.txt"}\NormalTok{, }\DataTypeTok{header=}\NormalTok{T, }\DataTypeTok{sep=}\StringTok{"}\CharTok{\textbackslash{}t}\StringTok{"}\NormalTok{)}
\KeywordTok{head}\NormalTok{(funseq)}
\end{Highlighting}
\end{Shaded}

\begin{verbatim}
##           is_annotated_in_encode is_sensitive is_ultrasensitive
## CR080767                       1            0                 0
## CR083996                       1            0                 0
## CR0911347                      1            0                 0
## CR0911356                      1            0                 0
## CR095246                       1            0                 0
## CR095443                       1            0                 0
##           is_motif_breaking target_gene_known taget_gene_is_hub
## CR080767                  0                 0                 0
## CR083996                  0                 1                 0
## CR0911347                 0                 0                 0
## CR0911356                 0                 0                 0
## CR095246                  0                 0                 0
## CR095443                  0                 0                 0
\end{verbatim}

\subsection{4 Build a classifier for prioritizing risk
varints}\label{build-a-classifier-for-prioritizing-risk-varints}

\subsubsection{4.1 Train a random forest
classifier}\label{train-a-random-forest-classifier}

Combining network, GWAVA features and FunSeq features, a random forest
model can be trained to predict risk SNPs.

\begin{Shaded}
\begin{Highlighting}[]
\CommentTok{#combine 3 types of features}
\NormalTok{group <-}\StringTok{ }\KeywordTok{as.character}\NormalTok{(}\KeywordTok{read.table}\NormalTok{(}\StringTok{"example_1/snp_labels.txt"}\NormalTok{)[,}\DecValTok{1}\NormalTok{])}
\NormalTok{features <-}\StringTok{ }\KeywordTok{data.frame}\NormalTok{(topoFeature, gwava, funseq, group)}
\NormalTok{RFmodel <-}\StringTok{ }\KeywordTok{trainMod}\NormalTok{(features)}\CommentTok{#train a random forest classifier}
\end{Highlighting}
\end{Shaded}

\begin{verbatim}
## [1] "The random forest model has been trained!"
\end{verbatim}

\subsubsection{4.2 Predict causal disease
variants}\label{predict-causal-disease-variants}

By providing the trained random forest model with feature values, users
can compute the prediction score for a list of snps or candidate
variants for being risk snps or non-risk snps.

\begin{Shaded}
\begin{Highlighting}[]
\NormalTok{prob <-}\StringTok{ }\KeywordTok{predMod}\NormalTok{(features[,}\OperatorTok{-}\KeywordTok{dim}\NormalTok{(features)[}\DecValTok{2}\NormalTok{]], RFmodel)}\CommentTok{#estimate the probablity score using trained model after removing the label/group informatin}
\KeywordTok{head}\NormalTok{(prob)}\CommentTok{#display the probablity score as positive or negative snps}
\end{Highlighting}
\end{Shaded}

\begin{verbatim}
##             neg   pos
## CR080767  0.442 0.558
## CR083996  0.386 0.614
## CR0911347 0.444 0.556
## CR0911356 0.402 0.598
## CR095246  0.432 0.568
## CR095443  0.420 0.580
\end{verbatim}


\end{document}