Meaning of 'tag' in the context of variants

Question

[andrewl]

I would like to understand what the meaning of 'tag is in the following paragraph. Does it mean "correlated with" or does it somehow act on the tagged gene?

The supplementary meta-analyses confirmed one additional independent locus that showed genome-wide significant association with CM (rs7412746 on chromosome 1q21; Amos et al., 2011), which was not included in our previous meta-analyses. Even though a number of independent signals were found to have a strong association with melanoma in certain loci (Table 1), some of these genes may not represent true susceptibility variants but may simply tag other genes within the same locus that have a more biologically plausible association with melanoma. For example, AFG3L1P and CDK10 genes on chromosome 16 and MHY7B on chromosome 20 most likely tag the well-known pigmentation genes of MC1R and ASIP, respectively (Bishop et al., 2009); e.g., the polymorphism rs258322 in CDK10 (16q24.3) is in linkage disequilibrium with the red hair allele of MC1R rs1805007, derived from data from the 1,000 Genomes (1KG) project (r2=0.58).

[Jean-Karim Heriche]

I interpret it to mean that the genes in question are picked up because they are on the same fragment of chromosome as the real causative gene(s).

[Ram]

A link to the paper where you read this would help.

Answer 1

[DG]

It is explained right in the text and answered by Jean-Karim above. The SNP that is being picked up as being associated with the phenotype is not causal but is in Linkage Disequilibrium with the true causal SNP. This is a pretty routine thing for GWAS analyses.