Inference_merip: Analysis MeRIP datasets with DESeq2.
In ZhenWei10/meripQC: Quality control and inference for m6A-seq datasets.

Inference_merip is a function to infer methylation and differential methylation given merip datasets.

1 2	Inference_merip(SE_M, MODE = "Meth", DM_METHOD = "DESeq2", PCA = FALSE, HDER = "Unknown", ROW_FILTER = 0, CQN = FALSE, GC_INDX = NULL)

`SE_M`	A `SummarizedExperiment` object with 1 necessary column in `colData` named "IP_input", its content should be a character vector consists of 2 values: c("IP", "input").
`MODE`	Could be either "Meth" or "DM", the later will conduct differential methylation analysis with the design:
`DM_METHOD`	A character string indicating the statistical method used in differential methylation analysis, could be one in c("DESeq2","QNB").
`PCA`	Wheather to save the PCA plot after rlog transformation, default is FALSE, the plot will not be generated when `DM_METHOD` = "QNB" while the `MODE` = "DM".
`HDER`	What should be the header of the PCA plot, applied when `PCA` = TRUE.
`CQN`	Wheather to normalize GC content dependency of methylation / differential methylation log2FC, default is FALSE.
`GC_INDX`	The GC content values for each features (rows) of the count matrix, it is required when CQN = TRUE. log2(Q) = intercept + I(Treated) + I(IP) + I(IP):I(Treated). The result is just the differences in conditioning effect, or the statistics for estimate before the term I(IP):I(Treated). We don't get this by contrast, because the information is already contained in coefficient estimate under the design above, and we don't need to linear combine (or a linear combination by t(c(0,0,0,1))) the estimates to get it. An additional column c("Perturbation") is necessary for this option, the Perturbation column has to include character "C" for control condition.