It is our belief that the genomic annotations and data formats will eventually converge into more standardized forms. At such time, an effective data retrieval confluence can be attained and elaborate annotation heuristics may not be necessary. This has begun to happen already with the LocusLink project where several databases are linked to provide a "one-stop-shopping" resource for genomic, genetic, and phenotypic information tied together by a single gene locus identifier. Overall, we were able to identify a number of clones that could serve as new cancer markers (whether they are suitable is a question to be answered in the wet lab). The moral of the story is that predictions of the end of science are rarely (if ever) correct. While some of the easier biological questions may have already been answered with the availability of the "complete" human genome, the genome contains many more answers for which we need to find the right questions. Focusing on the difficult, less understood areas is one way to ensure that we are vigilantly attempting to take advantage of the richness and depth of information available in the human genome.