%% \VignetteIndexEntry{Project Overview -- Slides}
%% \VignetteEngine{knitr::knitr}

\documentclass[xcolor=dvipsnames]{beamer}
\usepackage{BioconductorSlides}
\usepackage[round]{natbib}

\hypersetup{colorlinks,linkcolor=,urlcolor=Blue}
\AtBeginSection[]
{
  \begin{frame}<beamer>{Outline}
    \tableofcontents[currentsection]
  \end{frame}
}

<<cache-knitr, echo=FALSE>>=
knitr::opts_chunk$set(cache=TRUE)
@ 

\begin{document}

\title{Counting reads for RNA-seq}
\author{Martin T. Morgan \url{mtmorgan@fhcrc.org} \\ 
  Fred Hutchinson Cancer Research Center \\
  Seattle, WA, USA}
\date{25 August 2014}
\maketitle

\begin{frame}{Varieties of RNA-seq}
  \begin{enumerate}
  \item Known gene differential expression
    \begin{itemize}
    \item Genes, \Biocpkg{DESeq2}, \Biocpkg{edgeR}
    \item Transcripts
    \item Exons, \Biocpkg{DEXSeq}
    \end{itemize}
  \item Novel transcripts
  \end{enumerate}
\end{frame}

\begin{frame}{RNA-seq work flow}
  \begin{enumerate}
  \item Experimental design -- keep it simple; replicate.
  \item Wet-lab preparation -- covariates \& opportunities for `batch'
    effects
  \item Sequencing -- paired-end valuable for transcript-level
    inference
  \item Alignment -- typically to whole genome; requires aligner
    capable of gapped alignments
  \item Summary -- reads overlapping each gene or region of interest
  \item Analysis -- linear model (e.g., t-test) fit to each region
    of interest; `top table' of differentially expressed genes
  \item Comprehension -- annotation of differentially expressed
    regions, gene set enrichment, comparison to other studies,
    integration with other data types
  \end{enumerate}
\end{frame}

\begin{frame}{RNA-seq summary: counts per region of interest}
  \begin{itemize}
  \item Input: BAM files of aligned reads, typically one per sample
  \item How to count?
    \begin{itemize}
    \item What is an `overlap'?
    \item What (\Bioconductor) software to use?
    \end{itemize}
  \item Output: region $\times$ sample matrix of read counts
  \item \emph{Not} RPKM or other `normalized' measure
  \end{itemize}
\end{frame}

\begin{frame}{RNA-seq summary: how to count?}
  \begin{columns}
    \column{.5\textwidth}
    Counting modes\medskip\par
    \includegraphics[width=\textwidth]{figures/summarizeOverlaps-modes.pdf}
    \column{.5\textwidth}
    Counting in \Bioconductor{} 
    \begin{itemize}
    \item \Biocpkg{GenomicAlignments} \Rfunction{summarizeOverlaps()} --
      standard adn customized counting modes
    \item \Biocpkg{Rsubread} \Rfunction{featureCounts()} -- fast; Linux
      and Mac only
    \end{itemize}
  \end{columns}
\end{frame}

\end{document}