Specialized Case-Control Phenotype Misclassification Functions

Convenience functions for genotype chi-square case-control calculations in the presence of phenotype misclassification only. These focused functions are narrower than cc_mssn and cc_power: they use true affected and true unaffected genotype frequencies supplied directly by the user and do not apply locus heterogeneity or genotype misclassification.

cc_chisq_power_phenotype_misclassification(
  N_case,
  alpha = 0.05,
  g_aff,
  g_unaff,
  prev,
  theta = 0,
  phi = 0,
  k = 1
)

cc_chisq_mssn_phenotype_misclassification(
  target_power = 0.8,
  alpha = 0.05,
  g_aff,
  g_unaff,
  prev,
  theta = 0,
  phi = 0,
  k = 1
)

Arguments

N_case

Numeric \(> 0\). Number of cases for power calculations.

alpha

Numeric in \((0,1)\). Significance level.

g_aff

Numeric vector of true affected genotype frequencies, ordered consistently across affected and unaffected groups and summing to 1.

g_unaff

Numeric vector of true unaffected genotype frequencies, ordered consistently with g_aff and summing to 1.

prev

Numeric in \((0,1)\). Disease prevalence.

theta

Numeric in \([0,1)\). Probability that a truly affected individual is classified as a control, Pr(affected -> control).

phi

Numeric in \([0,1)\). Probability that a truly unaffected individual is classified as a case, Pr(unaffected -> case).

k

Numeric \(> 0\). Control-to-case ratio \(N_{ctrl}/N_{case}\).

target_power

Numeric in \((0,1)\). Desired target power for MSSN calculations.

Value

cc_chisq_power_phenotype_misclassification() returns a list containing sample sizes, phenotype-error parameters, observed case/control genotype frequencies, internal S, non-centrality parameter lambda, and power. cc_chisq_mssn_phenotype_misclassification() returns a list containing target power, phenotype-error parameters, observed case/control genotype frequencies, internal S, target non-centrality parameter lambda_star, and case/control/total MSSN.

Details

Phenotype misclassification is applied by mixing the true affected genotype distribution g_aff and true unaffected genotype distribution g_unaff using disease prevalence and the phenotype-error probabilities. Here theta = Pr(affected -> control) and phi = Pr(unaffected -> case). These specialized functions apply only phenotype misclassification; use the full case-control functions for combined phenotype misclassification, locus heterogeneity, genotype misclassification, or trend tests. The observed case distribution is a prevalence-weighted mixture of true affected and unaffected genotype probabilities conditional on being classified as a case; the control distribution is constructed analogously.

References

Edwards, B. J., Haynes, C., Levenstien, M. A., Finch, S. J., & Gordon, D. (2005). Power and sample size calculations in the presence of phenotype errors for case/control genetic association studies. BMC Genetics, 6, 18. doi:10.1186/1471-2156-6-18 . PMID: 15819990.

Gordon, D., Finch, S. J., & Kim, W. (2020). Heterogeneity in Statistical Genetics: How to Assess, Address, and Account for Mixtures in Association Studies. Springer. doi:10.1007/978-3-030-61121-7 .

See also

cc_power, cc_mssn, and case_control_genotype_misclassification.

Examples

g_aff <- c((1 - 0.05)^2, 2 * 0.05 * (1 - 0.05), 0.05^2)
g_unaff <- c((1 - 0.15)^2, 2 * 0.15 * (1 - 0.15), 0.15^2)

cc_chisq_power_phenotype_misclassification(
  N_case = 250, alpha = 0.01,
  g_aff = g_aff, g_unaff = g_unaff,
  prev = 0.05, theta = 0, phi = 0.01
)
#> $N_case
#> [1] 250
#> 
#> $N_ctrl
#> [1] 250
#> 
#> $alpha
#> [1] 0.01
#> 
#> $df
#> [1] 2
#> 
#> $prev
#> [1] 0.05
#> 
#> $theta
#> [1] 0
#> 
#> $phi
#> [1] 0.01
#> 
#> $g_case_obs
#> [1] 0.873760504 0.120546218 0.005693277
#> 
#> $g_ctrl_obs
#> [1] 0.7225 0.2550 0.0225
#> 
#> $S
#> [1] 0.07248965
#> 
#> $lambda
#> [1] 18.12241
#> 
#> $power
#> [1] 0.9134913
#> 

cc_chisq_mssn_phenotype_misclassification(
  target_power = 0.8, alpha = 0.01,
  g_aff = g_aff, g_unaff = g_unaff,
  prev = 0.05, theta = 0, phi = 0.01
)
#> $target_power
#> [1] 0.8
#> 
#> $alpha
#> [1] 0.01
#> 
#> $df
#> [1] 2
#> 
#> $prev
#> [1] 0.05
#> 
#> $theta
#> [1] 0
#> 
#> $phi
#> [1] 0.01
#> 
#> $g_case_obs
#> [1] 0.873760504 0.120546218 0.005693277
#> 
#> $g_ctrl_obs
#> [1] 0.7225 0.2550 0.0225
#> 
#> $S
#> [1] 0.07248965
#> 
#> $lambda_star
#> [1] 13.8807
#> 
#> $N_case
#> [1] 192
#> 
#> $N_ctrl
#> [1] 192
#> 
#> $N_total
#> [1] 384
#>