Infinium Methylation SNP

Question

[janhuang.cn]

I am using IlluminaHumanMethylation450kanno.ilmn12.hg19 package to obtain the SNP-associated probe information (probe_rs, probe_maf).

The reference manual says

In addition to the SNP information provided by Illumina, we have added independent information on the overlap of the 450k with various versions of dbSNP. The overlap is based on the mappings of the array to the hg19 genome provided by Illumina. As dbSNP we have used the ‘Common’ table from UCSC (ie. ‘snp137Common’).

I also found Infinium HD Methylation SNP List from the Illumina webpage

But I am still not sure how the SNP-associated probe (or overlap) is defined.

Would it be something like SNPs within xx bp of the CpG site?

Any information would be appreciated.

Answer 1

[andrew.j.skelton73]

If you're using Minfi (and by extension IlluminaHumanMethylation450kanno.ilmn12.hg19), there's 3 types of position that can be filtered with the dropLociWithSnps function, leveraging the snps parameter: SBE, CpG, and Probe. SBE is the single base extension, CpG is the actual CpG capture and Probe is the full capture sequence. Additionally you can use the maf parameter to threshold, which is usually around 0.05 (5%).

[janhuang.cn]

How is full capture sequence defined?

[andrew.j.skelton73]

The capture sequence is the genomic sequence that encapsulated the CpG site, and attaches to the oligonucleotide bead on the array.

[janhuang.cn]

Thank you very much

[jack]

thanks, what does maf 0.05 imply? why should these CpGs be removed?

[andrew.j.skelton73]

Minor Allele Frequency (MAF) - Measured differences in methylation could actually be a source of genetic variation in your population, thus if you want to minimise that bias then this function removes those probes.