Estimating prior and posterior values for methylation data that account for non-conversion rates is a time-consuming process. Significant increases in speed can be made by calculating in advance sets of data that will be re-used at several points of these analyses. This function populates the '@cellObservables' slot of a ‘countData’ object that contains a ‘nonconversion’ object in the ‘@sampleObservables’ slot.
A cutoff on the quantile (upper and lower) of the distribution on the non-conversion. Smaller values will give a marginal increase in accuracy at high computational cost. Large values will decrease accuracy somewhat but reduce the time needed for analysis. See Details.
For loci with large numbers of observed cytosines, the full dataset to be pre-computed will be very large. However, only the pre-computations near the average expression level will contribute significantly to the estimated priors and posteriors. The ‘tail’ parameter sets the quantile at which the distribution is considered to no longer contribute significantly to the results. Values below 0.1 are probably acceptable under nearly all circumstances.
countData object with the '@cellObservables' slot
populated with temporary values useful in the faster calculation of likelihoods.
Thomas J. Hardcastle
Add the following code to your website.
For more information on customizing the embed code, read Embedding Snippets.