The GPS information sent back includes longitude, latitude, time stamp, signal quality, etc.
We defined the mean STAMP signal density within these windows as the noise floor (NF).
We found a log-linear relationship between the STAMP signal and fractional methylation reported by the Illumina array.
To assess noise, we calculated a composite STAMP signal from the sequence tags within random 15 kb windows for all samples.
The STAMP signal had a broad dynamic range owing to its reporting regional methylation rather than fractional methylation at a single CpG.
We used this approach to calculate a STAMP signal surrounding all TSS, TTS, at CpGs interrogated by the Illumina HumanMethylation27 microarray and for 15,000 randomly selected genomic loci.
To calculate the STAMP signal, each tag was extended to a distribution of lengths modelling the DNA fragmentation pattern (Supporting Information S1).
Next, we compared the STAMP signal at 27,578 CpGs interrogated by direct bisulfite analysis using the Illumina HumanMethylation27 microarray (Supporting Information S1).
To assess the relationship between STAMP signal and CpG density, we separated probes into 11 bins based upon the CpG density surrounding the CpG interrogated by each probe.
To do this, we compared the STAMP signal in equally sized regions either 250 bp upstream or downstream of the TSS for each refSeq transcript.
