CHANGELOG.md

NGI-RNAseq

1.4 - 2018-03-26

The version 1.4 release contains major improvements to the NGI-RNAseq pipeline, especially focussed on the portability of the pipeline. Hopefully these changes will make it significantly easier for people to use the pipeline in different compute environments.

Many thanks to everyone who gave us feedback about the pipeline!

• Changed default config to base instead of uppmax
  • UPPMAX users now need to specify -profile uppmax
• Removed default reverse stranded config for UPPMAX profile
  • UPPMAX users now need to specify --reverse_stranded for the same behaviour
• Added an environment.yml file to easily create a (bio)conda environment for the pipeline
• Rewrote the Dockerfile to build the docker container using conda
• Switched UPPMAX configuration to use Singularity instead of environment modules
• Switched c3se Hebbe configuration to use Singularity & refactored config
• Added config profiles for two clusters at QBiC in Tübingen, Germany
• Made output from DupRadar, Biotype Counts and edgeR sample similarity use MultiQC Custom Content formatting
  • These results now show up in MultiQC reports for everyone, not just people with our custom MultiQC plugin
• Reorganised and rewrote much of the documentation
• Added a Troubleshooting section to the docs
• Fixed bug where BED12 generation failed when only GTF supplied.
• Changed dupradar script to use number of threads defined by nextflow process
• Fixed call to dupradar script that prevented interpretation of paired-ends reads and strand direction

1.3.1 - 2017-10-16

Hotfix to update version number in pipeline script.

1.3 - 2017-10-10

• Updated HISAT2 from v2.0.5 to v2.1.0
  • Uses --new-summary and --summary-file to give output that will work with MultiQC,
  • UPPMAX environment module load and Docker image
• Moved pipefail statement into config, applies to all processes
• Rewrote summary e-mail commands. Now uses sendmail
  • Proper multipart text/html e-mail with embedded images
  • If sendmail fails, falls back to sending plaintext using mail
• MultiQC process now runs using local executor for internet access on some UPPMAX clusters.
• UPPMAX featureCounts process now loads python/2.7.11 environment module
• New test script for uppmax with HISAT2
• Timeline and trace now always generated for every run
• Script now checks that the version of Nextflow is recent enough and warns if not
• New --help function to give usage help
• Software versions are now collected at run time and added to MultiQC and pipeline reports.
• RSeQC has been refactored, and geneBody_coverage.py moved into it's own process.
• The way config files work has been changed. Config settings are now inherited from base.config instead of uppmax.config
  • igenome.config needs to be last in the profile definition for the inhertence to work properly

1.2 - 2017-06-13

• MultiQC now runs using local instead of slurm with default config
  • Means that it stands a better chance of getting information from a remote database with plugins
• MultiQC now uses the workflow name for report title and filename if specified
• Fixed error where trimming parameters weren't set properly for --pico
• Made pipeline run in forward-stranded mode when using --pico

1.1 - 2017-05-23

New release of the pipeline with a few additions:

• Summary HTML report now created upon pipeline completion
  • This is saved in the results folder and also e-mailed if --email is set.
  • Nice way to alert the person running the pipeline that it completed and shows whether any errors were encountered.
  • Good for reproducibility as it logs all pipeline parameters to a easy to interpret file.
• Pipeline assumes that it's running with Paired-end files now
  • An error is now raised if the file glob doesn't give pairs of files
  • If running with single-end data, use the --singleEnd command-line option.
• Timelines and traces now created by default for the testing configs
• New configs and documentation about running the pipeline on AWS
• Made sure that the .bam files ended up in the main STAR directory when --saveAlignedIntermediates is used, instead of STAR/logs
• Made MultiQC load its config file through a channel instead of directly copying from baseDir

1.0.4 - 2017-04-21

RseQC hotfix, input file was not supplied properly to one of the scripts

1.0.3 - 2017-04-19

Hotfix to fix minor bug affecting strandedness for StringTie run.

1.0.2 - 2017-04-11

A couple of tweaks to help the pipeline in production:

• Trimming FastQ files and intermediate BAM files now not saved by default
  • This is configurable in the config or with --saveTrimmed / --saveAlignedIntermediates
• featureCounts merge process uses .collect() for better consistency

1.0.1 - 2017-04-10

This release includes a bugfix for the last major release relating to the strandedness of RSEQC.

• Single end reverse is now correctly +-,-+.
• Single end forward is now correctly ++, --
• PE forward is now correctly -1++,1--,2+-,2-+

1.0 - 2017-04-05

The pipeline has now been validated for use in our production work. This version includes some new features:

• The output from featureCounts is now merged into a single table and supplied along side the individual reports.
• markDuplicates JVM memory is now automatically scaled based on the process memory
• an html file with results documentation is now generated and supplied amongst the results
• It's now possible to configure the pipeline for different stranded libraries with just a simple CL flag.
• Additional support and documentation for other platforms than Uppmax. Inluding C3SE.
• • Numerous minor tweaks and improvements.

0.3 - 2016-12-13

In order to properly validate this pipeline and take it into production we need to tag a stable release. I've tagged specific software versions in the uppmax.config file.

0.2 - 2016-10-14

First (semi-) stable release of the new NGI-RNAseq pipeline, as we head towards deployment in production.