miscGlobal: Detect global shape variants in RNA-seq data
In hyochoi/Scissors:

Description Usage Arguments

This function discovers outlying subjects whose RNA-seq have abnormal shapes and provides the most outlying direction for each outlier.

1
2
3

miscGlobal(data, siglev = 1e-10, subt.mean = FALSE, PCnum = NULL,
  maxPCnum = 20, eps = NULL, rm.PCdir = TRUE, ADcutoff = 10,
  filter.dir = TRUE, less.return = TRUE)

`data`	precessed RNA-seq data from process_data
`siglev`	the significance level of the Chi-squared distribution. Default is 1e-10.
`subt.mean`	logical, whether to subtract mean before SVD. Default is FALSE.
`PCnum`	the number of PCs to be used. If NULL (the default) the number of PCs will be estimated.
`maxPCnum`	the maximum number of PCs to be used. Default is 20.
`eps`	tuning parameter for estimating the number of PCs. If NULL (default), eps = (1/nrow(data))
`rm.PCdir`	logical or numeric, the indices of PCs to be excluded from further study. If TRUE (the default), this automatically excludes a set of PCs that deviate from normality. If a sequence is specified, the corresponding PCs will be excluded.
`ADcutoff`	a cutoff value for checking the normality based on Anderson-Darling test statistic
`filter.dir`	logical, wether to filter out directions that deviate from normality. Default is TRUE.
`less.return`	logical, whether to show less results and not to return large matrices. Default is TRUE.