emi80 / docker-benchmarks

A benchmarks of genomic pipelines running in Docker containers

Docker benchmarks

This repository contains the benchmarks for these pipelines:

• Rnaseq-toy
• Nmdp-flow
• Piper-nf

Quick start

To reproduce the figures from raw data clone this repository:

https://github.com/cbcrg/docker-benchmarks.git && cd docker-benchmarks

Then pull the required Docker image:

docker pull cbcrg/make-peerj5515

Finally launch the make script using the following command:

docker run -v $PWD:$PWD -w $PWD cbcrg/make-peerj5515 bash make-figures.sh

(if you don't want or cannot use Docker, it will be required to download and install the required dependencies: R, Datamash, Make and Groovy. See the included Dockefile for configuration details).

It will create the following files:

./rnaseq-toy/table.txt
./rnaseq-toy/fig_correlation.png
./nmdp-flow/table.txt
./nmdp-flow/fig_correlation.png	
./piper-nf/table.txt
./piper-nf/fig_correlation.png

Structure of the repository

Benchmarks scripts and data are organised in separated folder depending the pipeline executed:

• Rnaseq-toy
• nmdp-flow
• piper-nf

Each of these folders contains a Makefile and other scripts used by it to generate the figures. Raw data of each executions are the files docker_{1..10}/trace.csv and native_{1..10}/trace.csv in the same folder.

The main files produced by the make process are:

• table.txt: the numbers included in the Table 1
• fig_correlation.png: the figure in included in the manuscript.

See the respective Makefiles for more details.

Replicate benchmarks

Pipeline executions can be replicated following the instructions in the respective repositories, shown below:

• nmdp-flow
• piper-nf
• rnaseq-toy