simple CNN without gene network prior

[bioops]

Hi,
Your work is very interesting!
Have you tried just simple CNN without gene network prior. It is like a 1D convolution with random order of genes. I wonder how your CNN model can be better than a traditional 1D CNN structure. Can you show some results, similar like figure 5 in your paper?
Thanks!