In 2003, the annual Text Retrieval Competition (TREC) had a Genomics track. The primary task in this track was to search through scientific abstracts - from a subset of the Medline collection - and find descriptions of the function of a given gene. At NRC, we made some modifications to our Litminer software and prepared submissions for TREC. Our results ranked very high among the 25 participating teams. In this talk, I will present details about the TREC task, about our software, and about the results. I will also discuss some of the other participants' efforts and give a quick overview of a secondary task in TREC-genomics.