RNAseq Multiple Read Group Input Support

RNAseq Multiple Read Groups

This release changes the input structure to allow for multiple read group inputs. Previously this workflow could only handle one read group. Please read the updated readme for instructions on how to input the reads files.

What's Changed

• 🔧 allow t1k ram adjustment by @dmiller15 in #58
• ✨ multi readgroup input by @dmiller15 in #59
• 🔖 bump to version 5.0.0 by @dmiller15 in #61

Full Changelog: v4.8.0...v5.0.0

T1k HLA Genotyping and STAR Diploid Beta

T1k HLA Genotyping

This release adds HLA Genotyping to the RNAseq workflow. The process is done using the T1k software. For more information, see our readme: https://github.com/kids-first/kf-rnaseq-workflow/blob/master/docs/T1K_README.md

STAR Diploid Beta

This release also adds a new beta workflow for STAR Diploid. It uses a patient's DNA variant calls to create a "personal genome" (PG) for improved alignment. For more information, see our readme: https://github.com/kids-first/kf-rnaseq-workflow/blob/master/docs/STAR_2.7.11b_DIPLOID.md

Other Changes

• Updated RNASeQC tool

What's Changed

• 🚀 STAR Diploid BETA WF by @migbro in #52
• 🎉 add T1k HLA typer by @dmiller15 in #54
• 🛠️ RNASeQC Updates by @migbro in #56
• 🔖 bump release to 4.8.0 by @dmiller15 in #57

Full Changelog: v4.7.0...v4.8.0

Quality of Life Updates to User Inputs

What's Changed

1. The user is no longer required to provide the input type. The workflow will now attempt to autodetect whether the reads are paired or single end. If the user wishes to override the autodetection, they can provide the is_paired_end input
2. The user is no longer required to provide kallisto_avg_frag_len and kallisto_std_dev for single end reads. The workflow will now attempt to autodetect the read length median and stddev values. If the user wishes to override the autodetection, they can provide the read_length_median and read_length_stddev values

• 🎨 simplify input variables by @dmiller15 in #49
• 🤖 add docker table gha by @dmiller15 in #50

Full Changelog: v4.6.0...v4.7.0

RNAseq and RMATs Input Autodetection

Autodetect Read Length and Strandedness

This update brings two quality of life improvements to our RNA workflows:

1. rMATs read length and variable read length inputs can be determined by the workflow and are no longer required. The rMATs workflow will now perform a polling of reads to determine these values. Whatever the most common read length found will be set to as the read length and if more than one read length is found variable read length will be set to true.
2. Strandedness can also be detected in most cases and as a result wf_strand_param is no longer required. This detection is accomplished through the use of RSeQC's infer_experiment script wrapped in a modified version of this script: https://github.com/signalbash/how_are_we_stranded_here.

Both of these improvements are optional and can be overridden by providing the inputs at workflow runtime. In the case of strandedness, if the script is unable to determine a strandedness, the user will need to provide wf_strand_param at input or the workflow will not be able to complete.

What's Changed

• 🤖 Compute rMATs readlength inputs by @dmiller15 in #47
• 🤖 automate strand detection by @dmiller15 in #48

Full Changelog: v4.5.0...v4.6.0

🐛 Critical Bug Fix - cutadapt mates to STAR handoff

If a user provided adapters for paired-end trimming, the workflow did not pass both reads (R1) and mates (R2) to STAR from cutdadapt, instead it passed only reads (R1) twice. If you ran the wf with bam input or fastqs with no adapter provided, or had a single end fastq with adapter provided YOU ARE OK. If you ran previous versions of V4 workflow with paired and fastq and provided adapters for R1 and R2 YOU MUST RE-RUN USING THIS WORKFLOW!!!

Other changes include:

• Making quality trimming optional
• Making variable read len for rMATs default per @naqvia recommendation
• Updates to docs

What's Changed

• 🐛 Critical bug fix by @migbro in #45

Full Changelog: v4.4.0...v4.5.0

🐛 annoFuse Reference Fix

• References in the annoFuse package were out of date in matching gene symbols for GENOCDE39+, causing annotations to be missed
• The package has been updated and the docker image pulls from this update
• A superfluous step was dropped and annoFuse format and run scripts brought into the repo and out of the docker
• User interaction with the wf will not have changed

What's Changed

• Feature/mb upgrade annofuse by @migbro in #44

Full Changelog: v4.3.0...v4.4.0

v4.3.0 ⚡️ Added more Efficient Default Reference

This release adds to the RNAseq wf a more efficient reference (800MB vs 30GB) for fusion annotator input. The annoFuse workflow simply has the default changed. For cavatica, also set a default instance for annoFuse. These combined changes reduce annoFuse run time from about 10-20 minutes to 1-3 minutes.

What's Changed

• 🏎 Updated reference for fusion annotator by @migbro in #42

Full Changelog: v4.2.2...v4.3.0

🧐 Upgrade dockers, annotate speed up

• Upgrade annoFuse docker, updated fusionannotator to pull from maintained image. No change in outputs
• Specify instance type for format fusion step - cuts down processing time by nearly half
• rMATS doc update for clarity

What's Changed

• ✏️ Edited doc for clarity by @migbro in #40
• ⬆️ Upgrade dockerpulls for annoFuse by @migbro in #41

Full Changelog: v4.2.1...v4.2.2

v4.2.1 Changed annoFuse Default Value

Updated annoFuse workflow to reflect software version updated and expected input formats

What's Changed

• 🔧 Change default fusion name column by @migbro in #39

Full Changelog: v4.2.0...v4.2.1

v4.2.0 Added Cram Input Support

✨ Cram Input Support

• Workflow can now convert cram to fastq, as well as bam/sam in order to properly align and process
• If you're uncertain the reference used in the input cram is completely registered (all contigs) in EBI, it is highly recommended you proved the reference fasta file used in the cram_reference slot, otherwise, reads that mapped to missing contigs may be dropped, resulting in incomplete analysis