I am trying to perform a GWAS performed with a custom Illumina array (MS Chip). I have an annotation data frame which contains IlluminaID (for example that column contains entries like this one: "exm-rs1003581-131_B_F_1990477648") another column which contain names (some are in this format "exm-rs1003582" and some are in this "exm-IND6-167673849"). Other columns contain chromosome, AddressID_A (integer), AddressID_B (which are mostly empty). On the other hand I have a gds file which contains SNPs in the format "exm2231095".
I am trying to use GWAStools to analyze this data. GWAStools requires for the AnnotationDataset a snpID variable which must be an integer, but that does no correspond with the SNP ID in the gds file and then it doesn't recognize the SNP ID. Does anyone know how to deal with this problem?
First, my apologies for not answering this sooner. The email notification feature of the support site appears to be broken, so I wasn't notified of your question.
GWASTools does assume that the snp.id node in a GDS file is an integer, and if you use the various conversion functions in the SNPRelate package (snpgdsBED2GDS, snpgdsVCF2GDS), that is what you will get. Since your GDS file has character values for snp.id currently, you can manually edit the file to store the character ids in a different node (snp.rs.id perhaps), and rewrite the 'snp.id node as an integer vector. (You will have to first delete the node and then rewrite it, since you are storing a different type.) This post has an example of how to use the functions in the gdsfmt package to directly modify a GDS file.
