Cuffdiff Bias towards positive fold changes

Question

[eclark28]

We recently ran some analysis using cuffdiff to perform differential gene expression testing. We have noticed something in our latest testing that is a bit odd. It appears cuffdiff favors genes that have a positive fold change over a negative fold change. We know that this is most likely not the case however, we are a bit unsure as we have never run into this issue. Our process is alignment to GRCH38 using HISAT2, DE with cuffdiff and GRCH38 gtf. Any ideas as to where this odd output could be coming from would be appreciated.