What should be included to make a good transcriptome/RNA-seq paper?

[poisson200]

Dear All,
I have done analyses on RNA-seq data that finds putative isoforms, differentially expressed genes/isoforms and gives a more complete picture of a tissues transcriptome.

I was wondering if anyone had any ideas on what is a "must have" inside a good transcriptome paper.

I have dabbled with bowtie/tophat/cufflinks/deseq,edgeR,DEGSeq and some others packages and now it is time to write something up.

It seems there are many options and things I could include out there and I would be interested in some views from other transcriptomers.

Thank you,

John.

[severin]

Biological questions

A paper should be centered around a biological question that you are interested in or is related to your research or grant. RNA-Seq is just one of many tools for conducting scientific research. The items that must be in a paper are dictated by what you are presenting in your paper. Experimental design, replicates, statistical analyses, table of statistics of your next gen data, the inherent assumptions you are making to simplify the analysis (what do you do with the multiple mapped reads?) the programs you are using to do the analysis (tophat, cufflinks). Pretty much everything that is needed for another group to reproduce your results.

[Joann]

More biology

Which public database will be deposited novel sequences? What are the gene ID numbers for the genes under discussion--the uniqur NCBI gene ID number inherently connotes species information. Including this information in the abstract will get your article more usefully text-mined, especially by sequence annotators.

[poisson200]

Thank you for your replies and advice.
Interesting what you say about NCBI gene id, as this is a number (if you mean Gene Entrez ID). I have not seen many manuscripts that report genes in abstracts this way. Though that could be showing how little I read.

As for biology. We are interested in certain tissues under certain conditions and how they react transcriptionally. I made the assumption about biology and was more aiming towards what should definitely be included bioinformatically.

No money for replicates, alas

Thank you.

[themerlin]

If you don't have money for RNA-seq replicates, make sure you have multiple qRT-PCR replicates.

[Joann]

Interesting what you say about NCBI gene id, as this is a number (if you mean Gene Entrez ID).

Text miners and annotators claim this number is gold, but seem reluctant to advance this insight to authors at large. At least one computationally big database outside NCBI I am familiar with uses it to normalize gene names across species and when authors place it in (open access) abstracts, this eliminates issues about copyright. As an experiment that can be replicated at any scale, I've performed manual sequence searches which save seconds or more of compute time when species info is included.