Detecting disease association with rare variants using weighted entropy.

The rapid development of sequencing technology and simultaneously the availability of large quantities of sequence data provide an unprecedented opportunity for researchers to conduct studies to detect rare variants associated with the disease. However, none of the current existing statistical methods has uniform power in all scenarios because they are more or less affected by nonfunctional variants and variants with opposite effects. Here, we present a robust approach to identify rare variants using weighted entropy theory. Here, this approach takes the proportion of the minor allele among all k variants as its probability distribution, which reduces the noise incurred by noncausal variants, and uses a weight to strike a balance between deleterious rare variants and protective rare variants, which makes our method impacted less by variants with opposite effect. Through simulation studies, we investigate the performance of our method for rare variant association analyses as well as for common variant association analyses and compared it with Burden test and the SKAT. Simulation studies show that the proposed method is valid and affected slightly by noncausal variants and opposite effect variants with high and stable power for various parameters set.