Thanks again for a great workshop! As promised, here's the GeneMANIA source code. The Lucene bits are in common/src/main/java/org/genemania/mediator/lucene. You'd probably want to have a look at LuceneMediator, which has most of the mapping code. It uses cglib's Enhancer class to defer the construction of the object graph until the appropriate getters are called.
If you want our data in plain text format, we have it hosted here: http://pages.genemania.org/data/
Please let me know if you have any questions. If you want to compile and run this, the provided READMEs give details about how to get started.