Skip to content

Latest commit

 

History

History
169 lines (105 loc) · 7.85 KB

README.md

File metadata and controls

169 lines (105 loc) · 7.85 KB
Raw

See links to instruction sections below:

The only thing you should need to edit is the config.yaml file (and DESeq2comparisons.yaml file if you are running DESeq2). All directory paths should end with a trailing /

How to make a sample sheet

The necessary columns are sample_name unit fast1 fast2 group exclude_sample_downstream_analysis

After that, you can include as many columns with sample metadata as you want.

What to put in each columns

sample_name - the name that you want each sample to have, for clarities sake I recommend using snake_case. Please no spaces, and don't start with a number, and don't use an periods in the sample name.

unit - this is for the situation when there are multiple fastq's per sample, such as as the following case

sample_name unit fast1 fast2 group exclude_sample_downstream_analysis
sample_one WTCHG_598911_202118 /SAN/vyplab/IoN_RNAseq/Bilal_Muscle_biopsies/fastqs/WTCHG_598911_202118_1.fastq.gz /SAN/vyplab/IoN_RNAseq/Bilal_Muscle_biopsies/fastqs/WTCHG_598911_202118_2.fastq.gz OPMD
sample_one WTCHG_606179_202118 /SAN/vyplab/IoN_RNAseq/Bilal_Muscle_biopsies/fastqs/WTCHG_606179_202118_1.fastq.gz /SAN/vyplab/IoN_RNAseq/Bilal_Muscle_biopsies/fastqs/WTCHG_606179_202118_2.fastq.gz OPMD
sample_two WTCHG_598911_203130 /SAN/vyplab/IoN_RNAseq/Bilal_Muscle_biopsies/fastqs/WTCHG_598911_203130_1.fastq.gz /SAN/vyplab/IoN_RNAseq/Bilal_Muscle_biopsies/fastqs/WTCHG_598911_203130_2.fastq.gz IBM
sample_two WTCHG_606179_203130 /SAN/vyplab/IoN_RNAseq/Bilal_Muscle_biopsies/fastqs/WTCHG_606179_203130_1.fastq.gz /SAN/vyplab/IoN_RNAseq/Bilal_Muscle_biopsies/fastqs/WTCHG_606179_203130_2.fastq.gz IBM
sample_two WTCHG_606180_203130 /SAN/vyplab/IoN_RNAseq/Bilal_Muscle_biopsies/fastqs/WTCHG_606180_203130_1.fastq.gz /SAN/vyplab/IoN_RNAseq/Bilal_Muscle_biopsies/fastqs/WTCHG_606180_203130_2.fastq.gz IBM

This is to allow trimming to occur on each fastq independently and then the merging is done later, if you don't have multiple fastqs per sample please just put a placeholder there. If downloaded from SRA, I recommend the accession, but just don't leave it empty, you could fill it with "placeholder".

sample_name unit fast1 fast2 group exclude_sample_downstream_analysis
axonal_control_1 SRR11430624 /SAN/vyplab/alb_projects/data/briese_tdp43_mouse_motorneuron/raw_data/SRR11430624.fastq axonal_control
axonal_control_2 SRR11430625 /SAN/vyplab/alb_projects/data/briese_tdp43_mouse_motorneuron/raw_data/SRR11430625.fastq axonal_control
axonal_control_3 SRR11430626 /SAN/vyplab/alb_projects/data/briese_tdp43_mouse_motorneuron/raw_data/SRR11430626.fastq axonal_control
axonal_control_4 SRR11430627 /SAN/vyplab/alb_projects/data/briese_tdp43_mouse_motorneuron/raw_data/SRR11430627.fastq axonal_control

fast1 fast2 - paths to the fast1 and fast2 files, if data is single end leave fast2 blank

group - a group name, useful for downstream analysis

exclude_sample_downstream_analysis - if a sample ends up being dirty or you don't want to analyze it, place a 1, otherwise leave blank

(Should be covered above - feel free to skip this paragraph...)

In the samples csv sheet, the unit column is used for when the fastq's for a single sample have been split. E.G if one sample has multiple fastqs this will tell you that. If you only have one fastq per sample, the unit column must not be empty. Please fill it out with either the sample name or just some place holder text like "a". It will not work if you just put a number.

(End of redundant paragraph)

How to run the pipeline

The pipeline has specific defined workflows. These are currently:

fastq_qc

  1. Trim reads with fastp
  2. Generate QC reports with FASTQC

interleave_fastq_qc

  1. Trim reads with fastp
  2. Generate QC reports with FASTQC
  3. Combine paired end read files into single, interleaved FASTQ file with bbmap

align

  1. Trim reads with fastp
  2. Generate QC reports with FASTQC
  3. Align reads to genome with STAR

salmon

  1. Trim reads with fastp
  2. Generate QC reports with FASTQC
  3. Quantify transcripts with Salmon

NB: salmon workflow currently only supports PAIRED-END READS

DESeq2

  1. Generate tx2gene mapping table for mapping genes to transcripts for DE from Salmon counts
  2. Run differential expression test with DESeq2

Minimal dependencies

You should have Snakemake executable from your PATH. Check that this is the case before submission by typing in "which snakemake" at the command line. It should say something like this: "/share/apps/python-3.7.2-shared/bin/snakemake""

To add Snakmake to your PATH. You'll also need to be using the correct version of Python. I recommend doing the following.

Open your .bash_profile: "nano ~/.bash_profile"

Add this line to your .bash_profile to source the Cluster Folk's file for setting your Python version

"source /share/apps/examples/source_files/python/python-3.7.2.source"

Close the .bash_profile, source it, "source ~/.bash_profile"

Confirm that snakemake is callable

"which snakemake" should now say "/share/apps/python-3.7.2-shared/bin/snakemake"

Submitting to cluster

Before submission I recommend that you test that everything looks correct with a dry run first. This can be done with:

snakemake -n -p -s workflows/{workflow}.smk

Submit to the cluster using the following command. The first argument should be your workflow of choice ({workflow}), followed by a specific run name for the job ({run_name}, optional but recommended).

source submit.sh {workflow} {run_name}

Testing and Updating the pipeline

If you make changes to the pipeline, be sure to test the pipeline works correctly before committing/submitting a pull request. I've made test files and datasets for paired end and single end runs, with corresponding downsampled FASTQs stored in our workspace on the cluster.

The configs and sample tables are ready to go, so to run the test datasets follow instructions below. If you perform a dry run and get a message like nothing to be done, go to the output test directory in vyplab_reference_genomes (you can find path in config/test_{type}_config.yaml) and delete the output. I recommend using align workflow, as this should include all possible rules in the pipeline.

Single end

Dry run

snakemake -n -p -s workflows/align.smk --configfile config/test_se_config.yaml

Submit to cluster

source submit_test_se.sh align {optional_run_name}

Paired end

Dry run

snakemake -s workflows/align.smk --configfile test_data_configs/test_pe_config.yaml -c 4 --use-conda

Run locally on an interactive node

qrsh -l tmem=16G,h_vmem=16G

snakemake -s workflows/align.smk --configfile test_data_configs/test_pe_config.yaml -c 4 --use-conda

What to do before pull requests or feature additions

Please make sure you have done the following before committing changes/submitting a pull request:

  • Ran pipeline (with test datasets) without workflow errors
  • Updated all config/*_config.yaml (i.e. including test examples) with new parameters
  • Updated README or config with additional instructions/comments (if adding functionality)

If you're adding a new workflow, also make sure to include a submit script.

(this is mostly a reminder for me...(Sam))

multiqc -p -o test_data_analyzed/paired_end/multiqc/align/ test_data_analyzed/paired_end/qc/fastqc/ test_data_analyzed/paired_end/fastp_trimmed/ test_data_analyzed/paired_end/STAR_aligned/ test_data_analyzed/paired_end/feature_counts/