---
title: "Multi-Tissue Analysis"
author: "Michael T. Zimmermann"
date: "`r Sys.Date()`"
output: rmarkdown::html_vignette
vignette: >
  %\VignetteIndexEntry{Enrichment Vignette}
  %\VignetteEngine{knitr::rmarkdown}
  %\VignetteEncoding{UTF-8}
---

In this vignette, we will analyze a gene expression dataset with samples from multiple tissues. We will:
*download a public dataset
*identify the genes expressed in two tissues
*run enrichment analysis, cognizant of each tissues' expression profile
*visualize network-based relationships between the tissues' expression profiles


# Enrichment to Identify Tissue-Specific Patterns

We will use data from BgeeDB normal-tissue expression. In research, we will typically want to compare normal to one or more treatment or disease groups. Thus, consider this as an illustrative example.

```{r bgee1, echo=TRUE, eval=FALSE, warning=FALSE, fig.width = 7, fig.height = 7, fig.align='center'}
library(RITANdata)
library(RITAN)
library(BgeeDB)
bgee <- Bgee$new(species = "Homo_sapiens", dataType = "rna_seq", release = "13.2")
data <- getData(bgee)

e <- formatData(bgee, data[[1]], callType = "present", stats = "rpkm")
# str(sampleNames(e))
# str(featureNames(e))
# str(phenoData(e))
# table(phenoData(e)@data$Anatomical.entity.name)

## -------------------- -
## Get expression in two tissues
tmp <- exprs(e)[ , phenoData(e)@data$Anatomical.entity.name == "heart" ]
i <- apply( tmp, 1, function(x){ any(is.na(x)) })
expr_heart <- tmp[ !i, ]

tmp <- exprs(e)[ , phenoData(e)@data$Anatomical.entity.name == "skeletal muscle tissue" ]
i <- apply( tmp, 1, function(x){ any(is.na(x)) })
expr_skele <- tmp[ !i, ]

library(venn)
venn::venn( list(Heart = rownames(expr_heart),
                 Skeletal = rownames(expr_skele) ),
            cexil= 1, cexsn = 1, zcolor = "style" )

## -------------------- -
library(biomaRt)
ensembl <- useMart("ensembl", dataset = "hsapiens_gene_ensembl" )
map_heart <- getBM( attributes=c('ensembl_gene_id','ensembl_transcript_id','hgnc_symbol'),
                    filters = 'ensembl_gene_id', values = rownames(expr_heart), mart = ensembl )
map_skele <- getBM( attributes=c('ensembl_gene_id','ensembl_transcript_id','hgnc_symbol'),
                    filters = 'ensembl_gene_id', values = rownames(expr_skele), mart = ensembl )

## -------------------- -
## Enrichment Within Each Tissue

to do...


## -------------------- -
## Network Interactions Within Each Tissue

to do...


## -------------------- -
## Similarities and Differences Between Tissues

to do...



```