Please help for RNA-seq data analysis using Cufflinks

[lewewoo]

Hi, I am a very fresh man in analyzing the RNA-seq results; my questions are may be too long, but I really appreciate any of the helps.
Basically, we are comparing gene expression changes and I decided to use Cufflinks pipline for this purpose. Please first help to go through our procedure:
1) RNA-seq data from Illumina;
2) TOPHAT alignment to produce SAM files;
3) Cufflinks package to process SAM to estimate the gene expression levels;

Secondly, I have questions about annotation reference file,
1) where I could get it? I heard it is GTF format, and I already got a file from UCSC table browser, and I chose total human all mRNA for my case, is this right way?
2) Cufflinks has the option using annotation reference, so I could use GTF file I got from UCSC?
3) I read some threads here, and it seems Cufflinks contains some reference file itself?
4)could I easily get expression results referred with Gene ID/Symbol? Should I use Cuffcompare or Cuffdiff for my purpose to compare expression changes of genes in different samples.

Millions millions thanks!

[kopi-o]

Regarding the annotation file:

1)-2) you can get a Cufflinks-usable annotation file from ftp://ftp.ensembl.org/pub/current_gtf/
4) yes, you will get gene symbols if they are in the annotation file. Use cuffdiff for looking at differential expression

[lewewoo]

Thanks!

To keep consistence, we are using hg18 for alignment, and I tried to find that in the link you provide, not good luck; so possibly, could i get the link to it...
thanks!

[kopi-o]

I see. I guess you probably want this one, then:

ftp://ftp.ensembl.org/pub/release-54/gtf/homo_sapiens/

(NCBI36 assembly = hg18)

[lewewoo]

Yes, I got V36 for that and also this... should be the same I guess; Thanks a lot!
BTW: Tomorrow is a big day for Royal Family UK, hehe

*****************************************
Hello Bogdan,

The best way to download the USCS genes table in GTF format would be to
use the table browser. After selecting your assembly of interest select:

group: Genes and Gene Prediction Tracks
track: UCSC Genes
table: knownGene
output format: "GTF - gene transfer format"

Then select "gzip compressed" and make sure you enter a name for the
output file (note that if you leave this field blank the browser will
try to return all of this information in your browser window and it may
stall or crash).


Regards,

Pauline Fujita
UCSC Genome Bioinformatics Group

http://genome.ucsc.edu