GEWIST (version 1.16.0)

effectPDF: Genetic interaction testing given effect size distribution

Description

Compute the optimal Variance Prioritization power and corresponding Levene's test p-value threshold for prioritization given the interaction effect size distribution using GEWIST.

Usage

effectPDF(distribution = c("beta", "normal", "uniform", "weibull"),
parameter1, parameter2 = NULL, parameter3 = NULL, p, N, theta_c, M, 
K = 20000, nb_incr = 50,  range = NULL, verbose = FALSE)

Arguments

distribution

distribution of interaction effect size. Possible distributions are:

"beta" for beta distribution "normal" for normal distribution "uniform" for uniform distribution "weibull" for weibull distribution

parameter1

the first parameter used in the corresponding distribution

parameter2

the second parameter used in the corresponding distribution, could set to be null

parameter3

the third parameter used in the corresponding distribution, could set to be null

p

minor allele frequency of the SNP, a number between 0 and 0.5

N

sample size

theta_c

proportion of quantitative trait variance explained by the covariate, should be a number between 0 and 1

M

total number of SNPs to be tested

K

number of GEWIST procedures, by default, set to be 20,000

nb_incr

number of effect size points in the range to be prioritized using GEWIST; by default set to be 50.

range

range of variance explained by interaction effect sizes, a vector of length 2

verbose

logical; if TRUE, for each interaction effect size, function returns a data.frame class object listing the VP power at each p-value, from 0.001 to 1 with 0.001 incremental increase.

Value

• A list with three components:
• Optimal_VP_power
  VP power to detect interactions at the optimal VP p-value threshold
• Conventional_power
  power to detect interactions without prioritization, i.e, VP power at Levene' test p-value of 1
• Optimal_pval_threshold
  levene'e test p-value at which optimal VP power is achieved

Warning

Computational time is directly proportional to nb_incr.

References

Deng W.Q, Pare G. (2011) A fast algorithm to optimize SNP prioritization for gene-gene and gene-environment interactions. Genetic Epidemiology. 35: 729-738. doi: 10.1002/gepi.20624

Pare G, Cook NR, Ridker PM, Chasman DI (2010) On the Use of Variance per Genotype as a Tool to Identify Quantitative Trait Interaction Effects: A Report from the Women's Genome Health Study. PLoS Genet 6(6): e1000981. doi:10.1371/journal.pgen.1000981

Levene H. (1960) Robust tests for equality of variances. In Contributions to Probability and Statistics: Essays in Honor of Harold Hotelling eds:I. Olkin, S.G. Ghurye, W. Hoeffding, W.G. Madow & H.B.Mann, pp.278-292. Stanford: Stanford University Press.

Examples

# Given a SNP with minor allele frequency of 10\% and a sample 
# of 10,000 individuals, we are interested in testing interactions 
# between this SNP and a covariate of effect size 10\%. The 
# total number of SNP is 500,000. Assume the unknown interaction 
# effect size has a Weibull distribution in the range of 0.05\% 
# and 0.3\% variance explained with 50 increments.Repeat GEWIST 
# for each of the 50 interaction effect sizes.  


library(GEWIST)
effectPDF(distribution = "weibull", parameter1 = 0.8, parameter2 = 0.3,
parameter3 = NULL, p = 0.1 ,N = 10000, theta_c = 0.1, M = 350000,
K = 20000, nb_incr = 50, range = c(0.05/100,0.3/100), verbose = FALSE)

## End of script