I am currently preprocessing data from the Illumina PsychArray chip. This chip includes normal tag SNPs and well as other genetic variants such as indels.

I intend to use the SNP data in a pathway analysis with the globaltest R package. The issue is that I have not dealt with indels before in the context of genomic data so I am unsure if I should include these variants in my analysis or not. Reasons for this include the variable length of indels and lack of genotype data (i.e SNP allele frequencies can be modeled in an addictive way, is this also true for indels ?).

Any assistance or opinions on this matter would be greatly appreciated!