cp4p (version 0.3.6)

adjust.p: Adjusted p-values for adaptive FDR control

Description

This function computes adjusted p-values for adaptive FDR control from a vector of raw (unadjusted) p-values.

Usage

adjust.p(p, pi0.method = 1, alpha = 0.05, nbins = 20, pz = 0.05)

Arguments

p

Numeric vector of raw p-values. Raw p-values are assumed without missing values, and between 0 and 1.

pi0.method

Numeric value between 0 and 1 corresponding to the proportion of true null hypotheses (non-differentially abundant proteins or peptides), or the name of an estimation method for this proportion among "st.boot", "st.spline", "langaas", "jiang", "histo", "pounds", "abh" or "slim" (see function estim.pi0 for details). The two-stage Benjamini and Hochberg procedure (Benjamini et al. (2006)) is also available according to an expected FDR given by the alpha parameter (write pi0.method="bky"). Default is 1 (classical Benjamini and Hochberg procedure (1995) is performed in this case).

alpha

A nominal type I error rate used for estimating the proportion of true null hypotheses (non-differentially abundant proteins or peptides) in the two-stage Benjamini and Hochberg procedure (used only if pi0.method="bky"). Default is 0.05.

nbins

Number of bins. Parameter used for the "jiang" and "histo" methods in estim.pi0. Default is 20.

pz

P-value threshold such as p-values below are associated to false null hypotheses. Used for the "slim" method in estim.pi0. Default is 0.05.

Value

A list composed of :

pi0

The proportion of true null hypotheses (non-differentially abundant proteins or peptides) used to adjust p-values.

adjp

A matrix of raw and adjusted p-values with rows corresponding to each test. First column corresponds to raw p-values and second column to adjusted p-values.

Details

The procedure uses an estimation of the proportion of true null hypotheses (non-differentially abundant proteins or peptides), the value or the name of which is precised in input. Next, this estimation is multiplied by the adjusted p-values of the Benjamini and Hochberg procedure (1995) to obtain the final adjusted p-values (see section 3 in Craiu and Sun (2008) for details).

The adjusted p-values of the Benjamini and Hochberg procedure (1995) and of the two-stage Benjamini and Hochberg procedure (Benjamini et al. (2006)) are computed using the R package multtest (Pollard et al. (2005)).

References

Y. Benjamini and Y. Hochberg. Controlling the false discovery rate: a practical and powerful approach to multiple testing. Journal of the Royal Statistical Society. Series B (Methodological), 289-300, 1995.

Y. Benjamini, A.M. Krieger, and D.Yekutieli. Adaptive linear step-up procedures that control the false discovery rate. Biometrika, 93(3):491-507, 2006.

R.V. Craiu and L. Sun. Choosing the lesser evil: trade-off between false discovery rate and non-discovery rate. Statistica Sinica, 18:861-879, 2008.

K.S. Pollard, S. Dudoit and M.J. van der Laan. Multiple Testing Procedures: R multtest Package and Applications to Genomics, in Bioinformatics and Computational Biology Solutions Using R and Bioconductor. Springer. 2005.

See Also

estim.pi0, calibration.plot

Examples

# NOT RUN {
#get p-values
data(LFQRatio2)
p=LFQRatio2[,7]

#adjust p-values by estimating the proportion of true null hypotheses 
#with the "pounds" method.
res_pounds=adjust.p(p, pi0.method = "pounds")

#proportion of true null hypotheses with the "pounds" method.
res_pounds$pi0

#plot ajusted p-values in function of raw p-values
plot(res_pounds$adjp)

#adjust p-values by estimating the proportion of true null hypotheses
#using the two-stage Benjamini and Hochberg procedure with a FDR of 0.1.
res_bky=adjust.p(p, pi0.method = "bky", alpha = 0.1)

#proportion of true null hypotheses with the two-stage BH procedure.
res_bky$pi0

#plot adjusted p-values in function of raw p-values
plot(res_bky$adjp)

#compare the two-stage Benjamini and Hochberg procedure 
#with the "pounds" method
plot(res_pounds$adjp[,2],res_bky$adjp[,2])

# }