HIDATA to LOWDATA status and now false p value Cuffdiff isoform_exp.diff

[jmwhitha]

Hello SeqAnswers community,

When I first tried to run cuffdiff, I found that I had many HIDATA status for my isoform_exp.diff. After running it with 10X the default max-bundle-frags, I didn't have any HIDATA status left. However, I have many LOWDATA status now.

It's very strange. I am looking at a particular gene right now that has 167095 count but has a LOWDATA status while the same gene in the other condition has a OK status (821 count). Before adjusting the max-bundle-frags, they both had HIDATA.

Clearly, there is differential expression, but cuffdiff isoform_exp.diff says "no".
test_id = TCONS_00000809
gene_id = XLOC_000295
gene = Doesn't matter
locus = Doesn't matter
sample_1 = Cond1
sample_2 = Cond2
status = LOWDATA
value_1 = 7012.28
value_2 = 1.07803
log2(fold_change) = -12.6673
test_stat = 0
p_value = 1
q_value = 1
significant = no

Why is it giving these high p and q values?

Thank you and God bless,
Jason

[jmwhitha]

I take that back, actually I changed the max-bundle-frags to 70X the default.

Please, if anyone knows why my p-values are defaulting to one and my test_stat is defaulting to 0, I would greatly appreciate the information.

Thank you!