The computational process of our system is real-time, which is different from databases such as HaploReg and RegulomeDB, where the function annotations are pre-computed and stored in the database in advance. Therefore, it can dynamically deal with the genetic variants input by users with maximum flexibility. Despite large computational burden in the background when LD is considered, our system can finish the job of a meta GWAS data set (thousands of variants with moderate GWAS significance, P < 1.0 × 10−5) within a few hours even with LD from the 1000 Genomes Project. It will be much quicker when using HapMap LD. To exploit the regulatory properties of personal genomics data, GWAS3D accepts VCF-like format and can evaluate the deleteriousness of rare/novel variation altering gene regulation associated with personalized trait.
