There are 14 columns in cuffdiff output

awk -F"\t" '{print $1"\t"$8"\t"$9"\t"$10}' gene_exp.diff | head 
test_id value_1 value_2 log2(fold_change)
A2M-AS1|144571|locus1of1|1      0.0190991       0.0447137       1.22722
A2MP1|3|locus1of1|1     0       0       0
A2M|2|locus1of1|1       0.0261932       0.0171011       -0.615099
A3GALT2|127550|locus1of1|1      0.0287783       0       -1.79769e+308
A4GNT|51146|locus1of1|1 0       0       0
AA06|100506677|locus1of1|1      0       0       0
AACSP1|729522|locus1of1|1       0       0.00822783      1.79769e+308
AADACL2-AS1|101928142|locus1of1|2       0       0       0
AADACL2|344752|locus1of1|1      0       0       0
...

the log2(fold_change) was computed by log2(value_2/value_1)

where “value_1” comes from the first “bam file” and “value_2” comes from the second “bam file” as the following command

cuffdiff  -o /somePath/ /somePath/transcripts.gtf value_1_firstBam.bam value_2_secondBam.bam