PMC:3216508 / 11612-13094 JSONTXT

{"project":"2_test","denotations":[{"id":"22355539-17060359-135382458","span":{"begin":321,"end":323},"obj":"17060359"},{"id":"22355539-17529973-135382459","span":{"begin":497,"end":499},"obj":"17529973"},{"id":"22355539-19474294-135382460","span":{"begin":843,"end":845},"obj":"19474294"}],"text":"Our approach of SNP selection is different from the traditional way of tagging SNPs, where SNPs were usually selected based only on the linkage disequilibrium in a genomic region. The SNP selection process focuses more on efficiency, which maximizes the coverage but minimizes the cost (the number of SNPs for genotyping)16. After genotyping, one can search for the functional SNPs with close LD to the tagged SNP that shows significant associations, as we did previously in breast cancer scanning17. However, there are disadvantages of this approach. (1) The LD is usually inferred from a limited number of previously genotyped data, such as HapMap. It might be different in the current studied population. (2) It ignores the functional role of the SNP, and many tagged SNPs are in the gene desert regions, and have little biological meanings1819. Our SNP selection approach focuses on the functional role of the SNPs, particularly in the promoter region of the gene that might affect the transcription factor binding sites. We found this simple approach very effective in identifying functional SNPs for our IL22 promoter study. Among the thirteen regulatory SNPs we selected, five (38.5%) showed significant association with TB susceptibility. This number is very high compared with non-regulatory SNPs, of which we selected seven SNPs, with none of them showing significant association. However, the effectiveness of this approach on other genes is unknown and yet to be tested."}