customize_percentile: Generate Predicted Quantiles for Age Trends

In ComBatFamQC: Comprehensive Batch Effect Diagnostics and Harmonization

Generate Predicted Quantiles for Age Trends

Description

This function computes predicted quantiles for a specified feature and demographic group based on a GAMLSS model. The function interpolates predictions over a range of ages while accounting for fixed covariates.

Usage

customize_percentile(age_list, feature, q = 0.75, s = "F")

Arguments

age_list	A list containing all ROIs' true volumes, age trend estimates, and the fitted GAMLSS model.
feature	A string specifying the feature of interest within the age_list.
q	A numeric value between 0 and 1 representing the quantile to predict (e.g., 0.5 for the median).
s	A string indicating the gender of the group for which the predictions are generated (e.g., "F" for female, "M" for male).

Details

This function uses a GAMLSS model to generate predictions for a specified quantile and demographic group. The predictions are computed over a sequence of ages (age_test) that spans the observed age range in the data. The function adjusts for fixed covariates such as icv by using their mean values from the input data.

Value

A data frame containing columns for age, quantile type, prediction, and sex.

Examples

sub_df <- age_df[,c("Volume_1", "age", "sex", "ICV_baseline")] |> na.omit()
colnames(sub_df) <- c("Volume_1", "age", "sex", "icv")
age_list <- list("Volume_1" = age_list_gen(sub_df = sub_df))
customize_percentile(age_list, feature = "Volume_1", q = 0.5, s = "F")

ComBatFamQC documentation built on April 4, 2025, 12:24 a.m.