Time required to complete a transcriptom to Genome mapping project

[the_august]

Hello All,
I'm a new comer to NGS. However, I have years of experience in other field of Bioinformatics and I'm well
versed in scripting/programming. Besides, I've experience in general sequence analysis.
I wish to align raw squence reads of a transcriptome to a already sequence genome of the same organism to look for
novel transcripts.
I plan to use TOPHAT. I've access to high performance Linux Cluster at my University. Being a novice I think I'll need some time to learn
about these tools, stumble on some steps and retry things.

I wish to know how long it might take for me to complete such a project?

Thanks

[Brian Bushnell]

The first part won't take long, though it depends on the size of your genome and number of reads. You simply index the genome (one command), then map the reads (another command). Alternately, if you are only interested in novel transcripts, you could map to the transcriptome first, requiring high identity, then only map the remaining unmapped reads to the genome, which will greatly enrich for novel splices and leave you with less data to wade through.

The analysis will take a lot longer; how long depends on how much you automate it and how big and well-annotated the genome is. There are some tools in the Tuxedo package designed for identifying novel transcripts but I have not found them to work well. I highly recommend IGV, though, for visualizing aligned reads to identify novel splices and transcripts. Incidentally, when mapping RNA-seq data to the genome to look for novel transcripts, I recommend BBMap over Tophat; it's faster and more sensitive.

[the_august]

Many thanks for your answer. I'll be working with Mouse genome and a particular mouse tissue transcriptome.

Thanks

[GenoMax]

Since you already have experience with command line/programming it should not take longer then a month to complete the whole project (probably including the time you will need to spend on coming up to speed with things).

That time estimate is for a modest number of samples (say < 100). If you are going to do thousand(s) then the one month estimate does not apply