QL.results: Obtain p-values and q-values using results from 'QL.fit'

Description Usage Arguments Value Author(s) References See Also Examples

View source: R/QL.results.R

Description

Obtain significance results for quasi-likelihood model fits to RNA-seq expression count data using the methods detailed in Lund, Nettleton, McCarthy, and Smyth (2012).

Usage

1
2
	
QL.results(fit,Dispersion="Deviance",spline.df=NULL,Plot=TRUE)

Arguments

fit

The list returned by the function QL.fit

Dispersion

Must be one of "Deviance" or "Pearson", specifying which type of estimator should be used for estimating quasi-likelihood dispersion parameter.

spline.df

Optional. User may specify the degrees of freedom to use when fitting a cubic spline to log-scale(estimated dispersion) versus the log(average count). Default uses cross-validation in sreg function to pick appropriate degrees of freedom.

Plot

logical. If TRUE, the estimated dispersion versus the average count are plotted on a log-scale with the corresponding cubic spline fit overlaid.

Value

list containing:

"P.values"

list of matrices providing p-values for the QL, QLShrink and QLSpline methods, respectively. The i^th column of each element of pvals corresponds to the hypothesis test assigned in the i^th row of test.mat.

"Q.values"

list of matrices providing q-values for the QL, QLShrink and QLSpline methods, respectively. The i^th column of each element of qvals corresponds to the hypothesis test assigned in the i^th row of test.mat. Q-values are computed using the methods of Nettleton et al. (2006) JABES 11, 337-356.

"F.stat"

list of matrices providing F-statistics for the QL, QLShrink and QLSpline methods, respectively. The i^th column of each element of F.stat corresponds to the hypothesis test assigned in the i^th row of test.mat.

"m0"

matrix providing estimated number of true null hypotheses for each test(arranged by row) under each of the three methods(arranged by column). m0 values are computed using the methods of Nettleton et al. (2006) JABES 11, 337-356.

"d0"

vector containing estimated additional denominator degrees of freedom gained from shrinking dispersion estimates in the QLShrink and QLSpline procedures, respectively.

Author(s)

Steve Lund lundsp@gmail.com

References

Lund, Nettleton, McCarthy and Smyth (2012) "Detecting differential expression in RNA-sequence data using quasi-likelihood with shrunken dispersion estimates" SAGMB, 11(5).

See Also

QL.fit, NBDev, mockRNASeqData

Examples

1
2
  
## see examples for QL.fit()

QuasiSeq documentation built on July 10, 2017, 1:05 a.m.