If you already have a visualization set up, double check that datasets you are trying to visualize are of a type that can be displayed against a reference database (generally, any datatype that is coordinate based). Next, make sure that each dataset has the same "database" assignment that the visualization is based on.
Please be aware that for any visualization, any SAM datasets you may have need to be converted to BAM before visualization and wiggle/bedGraph datasets should be compressed into BigWig format.
If you used a Custom reference genome, you can still visualize. Create a Custom Build (User -> Custom Build), then make sure that is assigned to all datasets.
More about Custom reference genomes and builds is in our wiki here:
http://wiki.galaxyproject.org/Learn/CustomGenomes (Full Details, including troubleshooting help)
The question was a bit vague in that many factors can lead to an issue with reference genomes matching properly between visualizations and datasets (and they must be an exact match). But I hope this information helps to solve your problem, Jen, Galaxy team