We employ several methods for functional sequence assessment and predict functional consequence of different alleles of a SNP. To reduce the number of false positive results, we perform the predictions in only the most probable genomic regions for each category of functional sequence site (such as the gene promoter region for TFBS or the 3′ UTR for microRNA-binding sites) and use phylogenetic footprint information to filter out non-conserved putative functional sequence. The SNP selection algorithm uses functional prediction results to prioritize LD tag SNP selection. These LD tag SNPs capture other unexamined SNPs in and around a gene, including SNPs with unknown or unpredicted functional consequences. The web utility options allow an investigator to choose prediction methods and assign weights to those predictions for study-specific SNP selection. Functional sequence prediction is a rapidly developing field. The web server structure allows rapid updates as better methods of functional prediction become available and it allows expansion to include predictions on other biologic functions.
