• Home
  • Rationale
  • AWG Runs
  • Contact Us
  • DCC Interactions
  • Nozzle
  • Documentation
  • Presentations
  • QualityControl
  • Dashboard-Analyses
  • FAQ
  • Dashboard-Stddata
  • Data Usage Policy
  • ProcessFlow
  • Download
  • MAF Dashboard
  • Mission Statement
  • Nomenclature
• HomeNew

Nozzle R Package

Nozzle is an R package for generation of reports in high-throughput data analysis pipelines. Nozzle reports are implemented in HTML, JavaScript, and Cascading Style Sheets (CSS), but developers do not need any knowledge of these technologies to work with Nozzle. Instead they can use a simple R API to design and implement powerful reports with advanced features such as foldable sections, zoomable figures, sortable tables, and supplementary information. Please cite our Bioinformatics paper if you are using Nozzle in your work.

Documentation

• Firehose Reports At-a-Glance (a visual overview of Nozzle report features)
• API Overview (current release)

Presentations

• TCGA GDAC Research Call (11 August 2011)
• TCGA Face-2-Face Meeting (27 April 2011)

Report Examples

• Preprint of the Nozzle Bioinformatics Applications Note as a Nozzle Report (demonstrates custom styles).
• Advanced demo rendered to HTML (use navigation in the top left corner to navigate to additional reports).
• Firehose Analysis reportsfrom 24 October 2012 (click on "Browse" in the "HTML" column). Here are some highlights:
  • Glioblastoma Multiforme Copy Number Analysis with GISTIC2: Demonstrates figures, tables and supplementary information.
  • Lung Squamous Cell Carcinoma Gene Mutation Analysis: Demonstrates figures in supplementary information.
  • Ovarian Serous Cystadenocarcinoma Putative microRNA Target Analysis: Demonstrates use of citations and input parameter lists in the Methods & Data section.
  • Uterine Corpus Endometrioid Carcinoma mRNA CNMF Clustering: Demonstrates (experimental) "Print" feature and "Report an Issue" button in the utility menu bar.

Code Examples

• A very basic example from the GitHub repository.
• The advanced demo requires download of code and additional files to generate the demo reports. Demonstrates all Nozzle features.
• A Nozzle report mimicking a journal paper preprint. Uses a custom style sheet.

Download and Installation

Nozzle is available on CRAN and can be installed directly from the R shell.

Archived Versions

For reference only. As of version 1.0-0 Nozzle should be obtained from CRAN. Please note that the releases before 1.0-0 do not contain any documentation within the R package.

GitHub Repository

The GitHub repository for Nozzle is at http://github.com/parklab/nozzle.

License

Nozzle is open source software and licensed under the LGPL version 2.1.

Contact

All questions related to Nozzle should be sent directly to Nils Gehlenborg (nils@hms.harvard.edu). Questions about Nozzle related to the Firehose Analysis pipeline should be directed to gdac@broadinstitute.org.

Nozzle & Firehose

Nozzle Reports for Monthly TCGA Firehose Analysis Runs

• Dashboard (public reports available starting January 2012)

Guidelines for Firehose Developers

• Report Guidelines for Firehose Pipeline Developers

Technical Notes on Integrating Nozzle into Firehose Modules

• The path to nozzle libraries can be passed into your R script using the special variable in your manifest/hydrant.deploy file: <nozzle.path> (evaluates to /xchip/tcga/Tools/Nozzle/v1.current). In other words, Firehose will take care of providing this value without the need to expose it to the task configuration (just like <libdir>).
• To load the libraries in R: require("Nozzle.R1", lib.loc=nozzlePath, quietly=TRUE); Where nozzlePath is a variable passed into the R script containing <nozzle.path>.
• All the resources referenced in your report must be in the top directory. For example, nesting of directories for figure image files will not display correctly in the Firehose UI.
• The report must be named nozzle.html. This is the default file name produced by writeReport(myReport). However, writeReport(myReport, filename="report.html") will not work. 
• In some cases, use of the <R> variable in the manifest/hydrant.deploy file has caused problems. When to use <R> and when to use a direct path to Rscript is currently unclear.