RNAseq Tutorial (with Reference Genome)

Beatriz-Soriano

To perform a RNAseq analisys with a reference genome, you have to follow the next steps:

Open RNAseq tool of the GPRO suite.
Go to Transcripts Protocols > Step-by-Step Mode > Tophat/Hisat & Cufflinks Protocol.
Do the quality analysis. To do this, go to Preprocessing > Quality Analysis > FASTQC. There, use your raw FASTQ files as input.
After the quality analysis, you have to perform the preprocessing step, so you have to click on Preprocessing > Trimming & Cleaning and select the software you want to use. If the quality analysis indicates you that there are adapters in your samples, you have to use CUTADAPT to remove them. Next, use PRINSEQ to preprocess your FASTQ files. Depending on the quality analysis, you have to use a set of parameters or another, but there are some options that are normally used such as -trim_ns_left, -trim_ns_right, -trim_qual_right, -min_len, -max_len, -min_qual_mean and -ns_max_p.
Once you have preprocessed your FASTQ files, it's time to map your reads against the reference genome using Tophat. To access to Tophat, click on Mapping > Tophat. Here, you have to provide an index file or a reference genome if you don't have it, your preprocessed FASTQ files and a GTF file with the annotation. It's recommendable to build a new transcriptome clicking on "Build New Transcriptome" option.
After you get the BAM files from the mapping step, you have to run Cufflinks. To access to Cufflinks go to Transcriptome Assembly > Cufflinks. Here, drag your BAM files to the corresponding dialog, introduce the GTF file containing the annotation and the reference genome inside two options: Reference sequences [seq_fasta] (Cuffmerge option) and -b option (frag-bias-correct [genome.fa]). If you only want to obtain the results for the known transcripts, click on "Do not Assemble novel transcripts".
Finally, you can perform the differential expression analysis using Cufflinks which is located in Differential Expression Test > Cuffdiff. You just have to supply the GTF file, your BAM files, sample group, the comparison you want to do and the reference genome in the -b option (Reference genome for bias correct).

For further information, you can consult the https://biotechvana.com/gpro/tool/rnaseq/manual/overview

tripleR

Dear B.
In my case I have to map my reads against mouse ref genome. Shall I create an index file, providing the fasta file from mouse genome (i.e., downloades from EMBL), or there is a easier way to incorporate this information (the same will apply, I guess) to use the GTF file once reads have been mapped

tripleR

In case I have several fastq files from different runs from the same sample, can I use them together or should I first merge them in a single file? Also, to map different samples, can I make the alignment in the same queue or it is better to make it one by one (ask me for a print screen if you need to explain better)

llorens

Hi there Ricardo,
to create the index file yes you should to use the fasta file from ensembl with the last genome release.
The GTF as well as the index should be used to perform the mapping step with top hat. Bea will guide you with the parameters etc.

llorens

if it is the same sample and same sequencer it is perhaps is easier yes if you merge first the fastq files into a single one and the use the merged one as query

tripleR

Is FastqCollpaser (within Preprocessing/PrepSeq) a tool to merge sequences?

tripleR

ESEMBL offers several alternatives of fasta files (DNA, cDNA, CDS, etc.). For a RNAseq analysis is better to use genomicDNA fasta file or will it suffice to use the cDNA fasta file?

Beatriz-Soriano

What FastqCollapser does is to remove duplicated reads. It is useful when you want to build a de novo assembly, but don't use it if you want to perform a RNAseq analysis because you will remove the differences of gene expression between samples.

tripleR

Understood, thanks!

Beatriz-Soriano

For a RNAseq analysis is better to use genomicDNA fasta file or will it suffice to use the cDNA fasta file?[/quote]

You have to use the genomic DNA fasta file, specifically that file which contain "toplevel" in its name. In your case, the file name is "Mus_musculus.GRCm38.dna_sm.toplevel.fa.gz".