In the hg18 directory there is a bigBed file, which only displays on chr7, that was created following the steps in http://genome.ucsc.edu/goldenPath/help/bigBed.html#Ex3
In summary the steps were to take a bed file, and when using bedToBigBed to add the -extaIndex option to include the name and geneSymb columns as information to index so that it can be searched. Then in the hg18/trackDb.txt a simple line was added, "searchIndex name,geneSymb", enabling the browser to search those fields when the hub is connected and browsing the bigBed file. For example, you can search MMD2, and since it is in the bigBed on chr7, it will show up as a possible result.
The searchIndex setting requires the input BED data to be case-senstive sorted (sort -k1,1 -k2,2n), where newer versions of the tool bedToBigBed (available here) are enhanced to catch improper input.
In the hg19 directory there is a bigBed file that is similarly created with the -extraIndex option when using the bedToBigBed step so that the "name" field can be searched.
However a large second step has been taken in the hg19 directory to associate each of those bidBed indexed names to a second created index file. Some notes are outlined in the hg19/trackDb.txt. In essence a program called ixIxx creates an additional"trix" index files. Then in the trackDb.txt, along with the "searchIndex name" line, a second "searchTrix outFile3.ix" line is added. Beyond just being able to search for named items like "SJ_96833_AA|GG", you can also search for related indexed text to those named items such as "2053" from the line used during creating the second index: "SJ_96833_AA|GG 996 + 2053 -1 996".
You can learn more about creating such search indexes for hubs by visiting these pages: