Project Summary Early childhood caries (ECC) is the most common chronic disease in preschool children in the United States. It has been shown to have a substantial heritability, but no consensus knowledge of the genetic variants associated with ECC exists. Existing genome-wide association studies of ECC are few and performed single-locus association mapping using logistic regression of caries affection status or count regression of DMFS/DMFT/DFS/DFT. Drawbacks of those approaches include potential misspecification of age effect, weak signal from single SNP, and the burden of multiple testing correction. The first drawback leads to incorrect sizes for the model-based association tests. The second and third drawbacks hinder the power of the association mapping. Multi-locus tests with age to ECC or the counting process of DMFS as phenotype can address the first and second drawbacks and alleviate the third one. Penalized variable selection with age to ECC as phenotype can address the first and third drawbacks. However, age to ECC and the counting process of DMFS are inevitably interval censored in caries research studies, as continuous monitoring of caries affection or severity is impossible. This makes the existing multi- locus survival tests and high-dimensional survival regressions not applicable to genetic association studies of ECC. The goal of this project is to develop novel methods for population- based and family-based genetic association analyses of survival outcomes, which address the above drawbacks and the interval censoring complexity, to dissect the genetic architecture of ECC. The specific aims are 1) to develop a suite of set-based genetic association and interaction tests with survival outcomes subject to interval censoring and possible left truncation, 2) to develop a suite of set-based genetic association and interaction tests with panel count outcomes, and 3) to develop a set of high-dimensional variable selection methods for interval censored and possibly left truncated data. The new methods will be programmed into R packages to be disseminated through the Comprehensive R Archive Network. Additionally, we will apply the methods to a dbGaP data set, Dental Caries: Whole Genome Association and Gene x Environment Studies (DC-GAGE), to illustrate their utility and discover subject matter knowledge. The successful completion of this project will address analytic challenges that impede ECC genetic research, and advance the statistical methodology development for population-based and family-based genetic association analyses of survival outcomes in general. The application of the new methods to the DC-GAGE data will provide new insights into the genetic etiology of ECC.