We conclude with several caveats. First, computational methods for sequencing-based GWAS are still under development3,7,22,24, whereas SNP-array based GWAS is a proven method that produces high quality data that can be analyzed using readily available computational tools. Second, sequencing data requires additional computational resources beyond what is necessary to analyze conventional GWAS as the analysis pipeline of sequencing data is typically more demanding than for genotyping data. Third, sequencing-based GWAS of the type described here does not involve sufficient coverage to discover rare variants and to associate them with disease; thus, as with SNP arrays, the power of this approach is limited to common and (to a lesser extent) low-frequency variants. Fourth, although results from our empirical IHCS sequencing data are encouraging, no study to date has used sequencing-based GWAS to identify new disease risk variants. A priority for future work should be to carry out studies that demonstrate that this approach can discover new associations between genetic variants and common diseases.
