The first assigned mRNA/gene that is listed in the annotation file for a given probe set or transcript cluster is considered the "best" assignment based on data source quality ranking and assignment scoring system. To get a single best mRNA/gene assignment you can simply take the first one in the list in the annotation file.
Sometimes, there might be multiple best hits between transcript clusters (or probe set) and a mRNA/gene, with identical score and data source quality. In that case, the transcript assignments are ranked based on their descriptions (e.g., title from the GenBank record).
For transcript clusters that detect genes that are alternatively spliced, the different spliced forms will be ranked by length with the longest ones listed first.
Find additional tips, troubleshooting help, and resources within our Microarray Analysis Support Center.