Comparing two RNAseq analysis tools

This repository is the final deliverable for our tutored project as a group of Master bioinformatics students.
The goal was to make and benchmark 3 pipelines using STAR+HTseq count, STAR quantmode, and Kallisto for RNAseq analysis. These were made with Snakemake.
The Snakefile contains all 3 pipelines. There is also a config file to allow easy customization of the pipeline.

Kallisto pipeline

It runs the 2 main Kallisto commands:

• The first one indexes the cDNA fasta file
• The second runs the quantification tool
• We added an extra transcript-to-gene tool to convert the counts to gene counts

STAR-HTseq pipeline

Here we have to use 3 different pieces of software :

• Star itself, to index the genome and map the reads on it
• Samtools, to convert the .sam output of Star into a .bam, and then to sort and index this file
• Htseq-count, which is the quantification tool for the reads we got using Star.

STAR quantmode pipeline

This pipeline uses the same index as the previous pipeline, followed by the built-in STAR quantmode command.

• Index the reference genome with STAR
• Count the reads by feature with STAR quantmode

Setting it up

• Place the Snakefile in a folder along with the config.json and the /input/ folder.

• Reads must be paired-end (2 files per condition), and placed in the /input/ folder.

• You must change the read extension in the config file, ex: "READ_EXTENSION" : ".fastq"

• You must change the paths for the DNA, cDNA and the GTF file in the config.json file, ending with a /

• The genomes must be .fa

• If the reads are compressed, change the SUP_STAR_COMPR parameter in the config file to "--readFilesCommand zcat", otherwise leave it empty: "".

• You must add the list of experimental conditions as a list in the config file, ex: "SAMPLE": ["C1reads", "C2reads"]

• You can also change the separator between the read name and number, ex: "C1reads_1" "C1reads_2" => "SEPARATOR":"_"

• To launch: move to your directory and type snakemake in the terminal, add -j [number of threads] to choose the total number of threads to be used. You can change the number of threads per rule in the config file: 3 rules can be run at the same time, so the total number needs to be three times the per rule thread number. Example:
  Console command: snakemake -j 12
  In the config file:"THREADS" : "4"