Based on bisulfite sequencing data from GSE46692, we have reconstructed the methylation profiles of Emiliania Huxleyi and Aureococcus anophagefferens. Both organisms have been shown to exhibit very clear periodic methylation patterns \cite{Huff_2014} and we are looking to replicate these results using our own pipeline of analysis.
We started with GFF files containing count data in the same manner as pictured in Fig. \ref{fig:hmm}b. After basic pre-processing, we ran the data in BSHMM to reconstruct a smoothed methylation profile instead of relying on fractional estimates. At this point, visualizing the results in IGV already gives a very strong impression of a repetitive pattern (Figure \ref{fig:acf1}, bottom) We go on by running an autocorrelation function over the methylation profile based on a full convolution of the signal with itself and we finally detect the size of the most meaningful period using a peak detection algorithm. What we have found is very consistent with the published findings, the periodic signal is strong and the periods are similar with a negligible difference of 5 to 10bp.