my understanding for cuffdiff output

[Huijuan]

Hi. I am a newbie in RNA-Seq analysis and now I am doing the pipeline with several software. I want to use cufflinks ,cuffcompare and cuffdiff to get DE genes,alternative splicing and novel transcripts and now I am confused about the output files cuffdiff gave me and I don't know what to do next. so the queston is:
1 Attached is a image I made to try to figure out the relationship of each file
I got from cuffdiff. But I don't know if it's right. so any advice will be
appreciated
2 If I'm going to analysis the DE,should I get the gene list and do subsequent
analysis like GO enrichment?
If I'm going to analysis alternative splicing difference between the samples.
How should I proceed? statistics the number of AS events, make a
histogram or something?
If I'm going to analysis the novel transcripts ,what am I supposed to do？
Thank you in advance for your advice.
Huijuan

[honey]

cufflink output

Huijuan,

You have very good understanding of RNA-seq work flow. I agree that Cufflinks/ Cuffdiff documentation and authors are always silent on this subject. This same question is being asked several time here in this forum and in other such discussion forums. I am now feeling that it was just one publication/s and no further assistance to fellow researcher. I used to use this work flow you are using. I am not an expert but can suggest that you look for unique Ids from differential expression files in combined GTF files. For gene level I have taken FPKM values and taken average for gene from tracking file. However best answer can come from Cole (who is now at Broad Institute) if he can bring is some insight what these millions of files mean and how to answer the question with which someone start RNA-seq experiment.