Pipeline for single nucelotide conversion detection on RNA-seq data

Description

This pipeline uses the Snakemake workflow manager to process RNA-seq data from nucleotide converted datasets such as SLAM-Seq, TimeLapse-seq or DART-seq, based on the detection of single nucleotide conversions for targeted RNA identification. The code is based on the HISAT-3N alignment tool within the Snakemake framework for automated and parallel running, and combined with different stategies for nucelotide-conversion oriented analysis.

Settings requierments

HISAT-3N

Since the HISAT-3N tool is not yet available as a conda package, the user has to download HISAT-3N scripts manually following the HISAT-3N installation steps and indicate the location of the hisat-3n folder as a variable within the config.yaml (see below).

As soon as a compatible installation is implemented we will add it to the pipeline.

Snakemake

To use this pipeline you would first need to create a Snakemake environment and and activate it. See more.

Getting started

1. Use this template to create your own project repo.

2. Create the config/sample_manifest.tsv using the following structure:

   1. Identifier: Unique sample ID used on the raw .fastq.gz sequencing files
   2. Sample_type: Tissue, cell type, etc.
   3. Treatment: Condition (treated/untreated, fed/starved, etc.)
   4. Bio_rep: Biological replicate number
   5. Labeling_method: Metabolic labeling method (SLAMseq or TimeLapse)
   6. Labeling_time: Incubation time for methabolic labeling
   7. Target_genome: ENCODE target genome for sequencing (supported options: M25, M32, 19, 38)
   8. Sequencer: Sequencer to define —2colour parameter for the trimming (supported options: HiSeq4000, NovaSeq, NextSeq500, NovaSeqX)
   9. Seq_mode: Library preparation strategy (supported options: mRNA, totalRNA)
   10. Seq_length: Sequencing length
   11. Fastq_handle: Particular handle useful for raw .fastq.gz files selection using the name from the sequencing facility (number, extension, etc.)
   12. Fastq_lanes: Number of .fastq.gz files lanes from the sequencing facility (used for merge_fq_lanes rule)

   An example of the sample manifest is available here.

3. Copy your raw sequencing data to resources/fastq_seq/raw, the names of the files must correspond to the values indicated on the sample manifest columns, using the following structure:

       # Read 1 - Lane 1
       {Identifier}_S{Fastq_handle}_L001_R1_001.fastq.gz
   
       # Read 1c
       {Identifier}_S{Fastq_handle}_L002_R1_001.fastq.gz
   
       # Read 2 - Lane 1
       {Identifier}_S{Fastq_handle}_L001_R2_001.fastq.gz
   
       # Read 2 - Lane 1
       {Identifier}_S{Fastq_handle}_L002_R2_001.fastq.gz
   

   The names inside the {} correspond to the column name on the config/sample_manifest.tsv. The pipeline autoamtically merges lane files (if applicable) and renames the files to:

       # Read 1
       {Sample_type}_{Treatment}_Chase-time_{chase_time_h}_Bio-rep_{Bio_rep}_R1.fastq.gz
   
       # Read 2
       {Sample_type}_{Treatment}_Chase-time_{chase_time_h}_Bio-rep_{Bio_rep}_R2.fastq.gz
   

   Again, with the corresponding values on the config/sample_manifest.tsv. If yopu wish to change the name of the raw files yo can do so changing the values of TARGETS['raw_seq'] on the Snakefile and merge_lanes.smk rule.

⚠️ Warning

Changing the structure `{Sample_type}_{Treatment}_Chase-time_{chase_time_h}_Bio-rep_{Bio_rep}` of filenames is not recommended and we don't ensure the proper functioning of the pipeline once it is done.

4. Check and change processing parameters on your config/config.yaml according to your analysis set-up.

5. Run the pipeline.

   Your pipeline should look like this:

   <YOUR PROJECT FOLDER>
       ├── config
       │   ├── config.yaml
       │   └── sample_manifest.tsv
       ├── profiles
       |   └── profile
       ├── resources
       |   └── fastq_seq
       |       └── raw
       |            ├── {Sample_type}_{Treatment}_Chase-time_{chase_time_h}_Bio-rep_{Bio_rep}_R1.fastq.gz
       |            └── {Sample_type}_{Treatment}_Chase-time_{chase_time_h}_Bio-rep_{Bio_rep}_R2.fastq.gz
       └── workflow
   

   Use the following command on your project folder:

       snakemake --profile profiles/<YOUR_PROFILE> -j<N_JOBS>
   

   For more options check the --help

Important notes

• This pipeline only works with paired-end sequencing reads.

• The default configurations for this workflow are suitable to run snakemake on the Helmholtz-Munich HPC and using conda to deal with software dependencies. If one wishes to run the pipeline on a different computing platform, the profiles need to be adapted accordingly.

• The pipeline only supports GENCODE genomes as reference, choose the right GENCODE Realease as value for the Target_genome entry in your config/sample_manifest.tsv. These are the equivalent realeases for GENCODE and UCSC databases:

  Organism	GENCODE Realease	GENCODE Genome	UCSC Genome
  Mouse	M25	GRCm38	mm10
  Mouse	M32 (Latest)	GRCm38	mm39
  Human	19	GRCh37	h19
  Human	38	GRCh38	h38

Authors and acknowledgment

Paulina Rosales-Becerra, Kevin Brokers & Robert Schneider

Contact

paulina.rosales@helmholtz-munich.de