what data normaly you use for analysis and how

Question

[Learner]

I don't have access to the raw files. I have access to the

htseq-count 
FPKM
FPKM-UQ

I have read all the questions asked in here. However, I still cannot find a good solution to the following questions

1- I want to find up and down-regulated genes, which package normally do you use and how do you convert them to the gene names?

2- which one of the above files are more reliable ? do you have any publication to refer to?

3- They come from TCGA, however, I cannot get the raw RNA seq and that is why I am trying to do the analysis based on one of the above files.

Thanks

[GenoMax]

If you can't or don't want to download raw data then you could use one of the TCGA data portals to get what you need. LinkedOmics is one option.

[Mo]

@genomax what is wrong with using htseq-count ? do you think it will not be a good idea to use that ? I personally think one can simply use that instead re-analysing the whole raw files (because you dont need to go trough splicing, alignment etc etc.

[GenoMax]

Sure that is the best option. There would still be some effort required on @Learner's part to use/analyze the raw counts. If the aim here is casual analysis then I was suggesting a web based alternative that will be simple(r) to use.

Answer 1

[WouterDeCoster]

Differential expression analysis is typically done using edgeR or DESeq2, for which you need the raw counts (from htseq-count). Have a look at this workflow to get started: Bioconductor RNA-seq workflow: gene-level exploratory analysis and differential expression.

and how do you convert them to the gene names?

Convert what to gene names?

[Learner]

@WouterDeCoster I am really sorry but this is not what I am looking for. There are 100 similar posts available. Please look at the right side of this post as an example. I am more searching to learn and not just ask random questions. So if you want a upvote, you should provide a solution or detailed answer. I appreciate the time you put but please consider that one answer should lead to a learning