Recount Website Public Datasets

Question

[narges]

Hi all,

For some comparisons I need to obtain the raw data of one of the public datasets available in Recount website and apply another method to obtain the count table (I want to apply the topHat over the raw dataset).

To have my comparison completely precise is it sufficient that I only download the fastq files of this dataset from the original publication and use them as the input of TopHat or are there other factors as well I should follow.

Thank you so much in advance.

Answer 1

[Michael]

I don't know this particular application but in general it is a good idea to use the same reference genome and genome build to keep the results comparable, if aligning from scratch. Then query and reference will be identical and only the alignment step and counting (probably) different. The genome annotations should also be kept comparable, so you should use the same annotation file as well, if your aim is comparability. If you want to have everything on the latest release instead, run it with the latest build and annotation though.

[narges]

Thanks. Do you mean the aligner by the genome build?

[Michael]

Nope, "genome build" refers to the version of the genome assembly used, like in GRCh37 vs. NCBI36 for the human genome.

see eg. http://www.ncbi.nlm.nih.gov/projects/genome/assembly/grc/human/