© 2013, Genentech

ReportingTools: an automated result processing toolkit for high throughput genomic analyses

Jessica Larson, PhD

Computational Biologist

Genentech, Inc.

larson.jessica@gene.com

July 19 2013

Outline

(1) ReportingTools introduction and basics

(2) ReportingTools and microarray experiments

(3) ReportingTools and RNA-seq experiments

(4) ReportingTools and knitr

(5) ReportingTools and shiny

2

ReportingTools

3

Example output and code from the package vignettes

http://research-pub.gene.com/ReportingTools/

4

Publishing a data.frame

5

There are three basic calls to publish to an HTML page:

(1)Define the report with

htmlRep <-

HTMLReport(shortName ='my_html_file',

reportDirectory = './reports’)

(2) Fill the report with

publish(iris, htmlRep)

(3) Close the report with

finish(htmlRep)

Publishing the Iris data

6

library(ReportingTools)

data(iris)

htmlRep <- HTMLReport(shortName = 'my_html_file',

reportDirectory = './reports')

publish(iris, htmlRep)

finish(htmlRep)

More basics with the Iris data

I. Adding plots (including basic, lattice, .png, .pdf)

II. Adding text (including links to other pages)

III. Adding additional tables

IV. Modify data frames upon publication (.modifyDF)

V. Publishing to CSV files (CSVFile)

7

Using .modifyDF

8

htmlRep <- HTMLReport(shortName = 'my_html_file_modify', title = 'Manipulating the data frame directly before

publishing’, reportDirectory = './reports')publish(iris, htmlRep, numdigits = 1, .modifyDF = list(roundLength, makeImages, addSpeciesLink, cleanUpDF))finish(htmlRep)

Outline

(1) ReportingTools introduction and basics

(2) ReportingTools and microarray experiments

(3) ReportingTools and RNA-seq experiments

(4) ReportingTools and knitr

(5) ReportingTools and shiny

9

Microarray examples

I. Publish output from limma (add new images and links)

II. Publish GO and PFAM analysis output

III. Publish gene sets

IV. Create index pages

10

Outline

(1) ReportingTools introduction and basics

(2) ReportingTools and microarray experiments

(3) ReportingTools and RNA-seq experiments

(4) ReportingTools and knitr

(5) ReportingTools and shiny

11

RNA-seq

Methods for:

I. edgeR exact tests

II. edgeR LRTs

III.DESeq

IV.DESeq2 (with .modifyDF and .toDF)

12

Outline

(1) ReportingTools introduction and basics

(2) ReportingTools and microarray experiments

(3) ReportingTools and RNA-seq experiments

(4) ReportingTools and knitr

(5) ReportingTools and shiny

13

knitr

Can easily integrate ReporingTools tables into knitr reports

Need to call ‘knit2html’ from report directory and specify the knitrHandlers in HTMLReport()

14

> setwd("reports")> library(knitr)> knit2html("/home/ubuntu/4BioCknitr.Rmd”)

Updates to run the knitr example (due to permission issues)

15

(1) Open 4BioCknitr.Rmd

(2) Save to the /home/ubuntu directory

(3) setwd(“reports”)

(4) Then call knit2html("/home/ubuntu/4BioCknitr.Rmd")

Outline

(1) ReportingTools introduction and basics

(2) ReportingTools and microarray experiments

(3) ReportingTools and RNA-seq experiments

(4) ReportingTools and knitr

(5) ReportingTools and shiny

16

Shiny and ReportingTools

17

Updates to run the shiny example (due to permission issues)

18

(1) Open server.R and ui.R

(2) Save these to the /home/ubuntu/reports directory

(3) setwd(“reports”)

(4) Then call myRunApp()

Changes to server.R

(1) Create a report with shinyHandlers to stream the HTML form of any elements added to our Report:

htmlrep <- HTMLReport(reportDirectory = "./", shortName="bigtest", handlers = shinyHandlers)

(2) Define a custom rendering function so that shiny can ‘hear’ elements being added to our report and insert them into the dynamic HTML:

renderRepTools <- function(expr, env=parent.frame(), quoted=FALSE) { func <- exprToFunction(expr, env, quoted) function(){ paste(capture.output(func()), collapse="\n") }

}

(3) Publish elements to our report within the expression passed to renderRepTools:

output$view2 <- renderRepTools({ publish(datasetInput(), htmlrep) })

19

Changes to ui.R

(1) Include the JavaScript and CSS files so that the ReportingTools tables function properly via custHeaderPanel function:

##this function accepts the tile and window title arguments by shiny's headerpanel function##plus the js and cs argumentscustHeaderPanel = function(title, windowTitle =title, js= NULL, css=NULL){

mytlist = c(lapply(js, function(x) tags$script(HTML(paste(readLines(x), collapse="\n")))), lapply(css, function(x) tags$style(HTML(paste(readLines(x), collapse="\n"))))) tagList(tag("head",mytlist), div(class = "span12", style = "padding: 10px 0px;", h1(title)))

}

# Define UI for dataset viewer applicationshinyUI(pageWithSidebar(

custHeaderPanel("ReportingTools", js = list.files(system.file("extdata/jslib", package="ReportingTools"), full.names=TRUE), css = list.files(system.file("extdata/csslib", package="ReportingTools"), pattern="bootstrap.css", full.names=TRUE), ),

20

Changes to ui.R

(2) Declare elements formated by ReportingTools as htmlOutput

mainPanel( verbatimTextOutput("summary"), htmlOutput("view2") )

This indicates to the shiny system that the output with be HTML code

21

Future methods

DESeq2 methods

Return the decorated DF

More flexibility with annotations

22

Acknowledgements

Jason Hackney

Josh Kaminker

Melanie Huntley

Christina Chaivorapol

Gabriel Becker

Michael Lawrence

Robert Gentleman

Martin Morgan

Dan Tenenbaum

23