identifying indels/novel fusions in somatic tumor/normal transcriptome RNA-seq Illumina data

Question

[genya35]

Hello,

I’m a researcher trying to figure out the best way to analyze tumor + normal somatic transcriptome RNA-Seq fastq files obtained from Illumina by using tools on Galaxy. The goal is to discover and visualize novel fusions and mutation. I’m hoping to get advice from users here that have done similar analysis.

Here is what I’ve done so far:

1. Ran FastQC
2. Ran RNA-STAR to mapping to reference genome(hg19) with gtf(all chromosomes) with all default parameters.
3. Ran Cufflinks on produced bam with the same gtf files. The output consisted of gene and transcript expression files.

Where do I go from here? Which files are the most helpful for identifying novel fusions? Do I filter the files to look only at transcripts/gene names that include Cuff* in the name? Should I also filter by highest PFKM values? How do I visualize/analyze normal and tumor output? How do I get a list of mutations (snvs,indels) present in the tumor files?

Should I use additional software for further analysis?

I would really appreciate a word of advice.

Thank you.

Answer 1

[Kevin Blighe]

If you're using Galaxy and STAR, then use STAR-Fusion for identifying the potential fusion transcripts.

There is a tutorial specifically for this on here: Galaxy Workflow ' EGA VCaP RNA-Seq Fusion gene detection'