venn_diagram_example.Rmd

---
title: "pompe-venn_diagrams_differential_express_pathways"
output: pdf_document
---

```{r setup, include=FALSE}
knitr::opts_chunk$set(echo = TRUE)


```

## R Markdown
```{r libraries, echo=F, message=FALSE, warning=FALSE}
require(readxl) #for reading in excel file
require(xlsx) #only needed for writing final excel file
require(venneuler)
require(VennDiagram) #another option for venn diagram 
require(dplyr)
require(tidyverse)
require(RColorBrewer)
```

```{r path info, echo=FALSE}
path <- "path/to/file"
file <- 'DifferentialExpression.xlsx'

filename <- paste0(path,file)
setwd(path)
sheets <- excel_sheets(filename)

AW <- read_excel(filename, sheet='AAV.vs.WT')
KW <- read_excel(filename, sheet='KO.vs.WT')
AK <- read_excel(filename, sheet='AAV.vs.KO')


dlist <- list(AW,KW,AK) # list of dataframes, more convenient for applying operations to all sheets

```



```{r  filtering data based on thresholds, echo= F}
pthresh <- 0.05 #cutoff for padj
thresh <- 1 #cutoff for log2 FC

#column 4= adj. log2 FC, column 8: adj. pval
dlist.trim <- lapply(dlist, function(x)
  filter(x,x[['padj']]<pthresh))
dlist.trim <- lapply(dlist.trim, function(x)  filter(x,abs(x[4])>=thresh))

AW.T <- dlist.trim[[1]]
KW.T<- dlist.trim[[2]]
AK.T<- dlist.trim[[3]]
```

genes matched based on geneid

color options from rcolorbrewer:
https://www.datanovia.com/en/blog/the-a-z-of-rcolorbrewer-palette/http://www.sthda.com/english/wiki/colors-in-r

```{r find overlapping genes for venn diagram incoporating directional changes}

#confirm not only geneid, but also direction of change
sign_add= function(x,stc=6,gn=2)
{  #receives row of a dataframe
  #returns genename w/ sign
  #stc =stat col (6), gn= gene name col (2)
  return(ifelse( sign( as.numeric(x[stc]) )==1,
                 paste0("+",x[gn]),
                 paste0("-",x[gn])
  ))
  
}

#used for matching and finding intersection for Venn diagram
KW_signed <- apply(KW.T,MARGIN = 1, sign_add, 6,'GeneName')
#reverse AK signs bc they need  to be opposite signs for matching
AK.T <- mutate(AK.T,revstat= -1*stat )
AK_signed <- apply(AK.T,MARGIN = 1, sign_add, stc=17,gn="GeneName")
```


```{r venndiagram package}
require(VennDiagram)
require(wesanderson)


require(RColorBrewer)
#COLOR options
# myCol <- terrain.colors(2)
# myCol <- c(rgb(66,105,160, max=255), rgb(.36,.57,.28))
# myCol <- wes_palette(name='Darjeeling2', n=2)
myCol <- brewer.pal(3, 'Dark2')#min is 3, just select the first 2 after
myCol <- myCol[1:2]


venn.diagram(
  x = list(KW_signed, AK_signed),
  category.names = c("KO/WT",'Treated/KO'), #Titles
  filename = '/destination/picture.png',
  output=TRUE,
  lwd=2,
  col=myCol, #border color
  fill=alpha(myCol,.6),
  # Output features
  imagetype="png" ,
  height = 2.2 , 
  width = 2.2 , 
  units='in',
  cat.just=list(c(0,-2) , c(1,-3)), #adjust position of titles
  #main.fontfamily = 'sans',
  resolution = 300)



```
sans=arial
serif= times new roman

https://rdrr.io/bioc/GeneOverlap/man/GeneOverlap.html
https://stats.stackexchange.com/questions/16247/calculating-the-probability-of-gene-list-overlap-between-an-rna-seq-and-a-chip-c
b<=a
a: number of genes in A
b:number of genes in B
n: total # of genes
t: intersec of A and B

dhyper(t, a, n - a, b) 
sum(dhyper(t:b, a, n - a, b)) # prob of seeing t or greater at the intersection
```{r hypergeometric venn diagram statistical test}
require(GeneOverlap)
x= 1006#intersection
m= 1238+1006#total genes in A
n= 32e3#total genes in Mm or Hs GENOME
k= 482+1006#total genes in B

phyper(x-1,m,n,k,lower.tail = F)


```