superSeq: Apply superSeq model to subsampling data
In StoreyLab/superSeq: Determining sufficient read depth in RNA-Seq experiments

Description Usage Arguments Value See Also Examples

The superSeq function fits a non-linear least squares model to subsampling data to learn the relationship between statistical power and read depth.

1 2	superSeq(object, control = defaultNLSControl, starts = NULL, new_p = NULL)

`object`	A subSeq summary object.
`control`	Specify convergence criteria for non-linear least squares algorithm. See `defaultNLSControl`.
`starts`	A list of combinations of starting guesses to try with `nls`. If NULL, uses `defaultNLSStarts`.
`new_p`	A vector of subsampling proportions to predict using the superSeq model fits. By default, triple the subsampling depth is predicted.

A superSeq object, which is a data.frame:

`fits`	The fitted objected from nls function
`subsample`	The subsampled object from subSeq
`predictions`	A data frame with the following columns: method used, total predicted discoveries, proportion read depth, predicted number of DE genes, and estimated statistical power.

fitnls

## Not run: 
library(superSeq)
library(subSeq)
library(Biobase)
# Load bottomly data
data(bottomly)
bottomly_counts <- exprs(bottomly)
bottomly_design <- pData(bottomly)
bottomly_counts <- bottomly_counts[rowSums(bottomly_counts) >= 10, ]
bottomly_proportions <- 10 ^ seq(-2, 0, 0.1)
# Apply subsampling methodology subSeq
ss = subsample(counts = bottomly_counts,
              proportions = bottomly_proportions,
              treatment=bottomly_design$strain, 
              method=c("voomLimma"),
              replications = 3,
              seed = 12345)
              ss_sum <- summary(ss)
              
# apply superSeq model
ss_obj <- superSeq(ss_sum)

# plot results
plot(ss_obj)

# summarise results
summary(ss_obj)

## End(Not run)