1. Files Needed

1. Gene annotation GTF
2. Aligned and sorted BAM files from tophat2

2. General Format

cuffdiff options genes.gtf condition1.bam condition2.bam

3. Run cuffdiff

cd ~/mydemo

cuffdiff --no-update-check -o cuffdiff_out –L wt,deletion mm9_chr1.gtf tophat_wt/accepted_hits.bam tophat_deletion/accepted_hits.bam

4. Parameters

-o	Directory to write all results
-L	Sample names separated by commas
-p	Number of processors to be used
--no-update-check	Do not check for new version. If not used, can  cause cuffdiff to hang/freeze.
-b	Correct for fragment bias to improve expression estimates
-u	Correct for multi-mapping reads

    

5. cuffdiff output

a. Cuffdiff creates ~20 output files
b. Log files are named *.info
run.info – command used to run cuffdiff
Normalized expression values are found in *.fpkm_tracking files
Gene level → genes.fpkm_tracking
Isoform level → isoforms.fpkm_tracking
Tab-delimited files can be opened in excel

c. Expression level differences are found in *.diff files
Gene level  → gene_exp.diff
Isoform level → isoform_exp.diff

d. gene_exp.diff - Important columns

q-value: p-value adjusted for multiple testing
Log2(fold change): fold-difference between condition 1 & 2