We wanted to simulate some actual elements and problems of real data collection, thus our decision to distribute the fine-mapping data, which consisted of 10,000 SNPs over the genome, from a specially-designed web site that limited the number of SNPs that could be requested to an approximation of what one could actually afford in a real research study, using the technology available in 2003. We packaged the chromosomal regions into groups of 20 contiguous SNPs (packets) and allowed participants to download a maximum of 20 packets. Each packet contained files for all replicates from all the populations studied, plus the same chromosomal region from a control population for doing case-control studies. The idea was that participants who did not request answers would first perform linkage analysis on the original data and then focus on those areas that produced the strongest evidence for linkage. Those who did request the answers, we assumed, would also concentrate on those areas in which there were disease-related loci.
