The Ochoa Lab is part of the Center for Statistical Genetics and Genomics at Duke University. We study evolution and disease under arbitrary population structure. Classical results often assume no population structure or a few discrete and independent subpopulations, which real data do not satisfy. We aim to improve:

  • Measures of genetic differentiation and population structure
  • Control of population structure in genome-wide association studies
  • Estimation of heritability
  • Detection of genetic loci under selection
  • Inference of genetic admixture

Important targets of study include recently-admixed individuals, such as Hispanics and African-Americans, who derive their ancestry from Sub-Saharan Africa, the Americas, Europe, and other places. Recently-admixed individuals vary dramatically in their ancestry composition, even among relatives, so their population structure (genetic correlation matrix) can be remarkably complex. Accurate modeling of population structure is necessary for studies involving these individuals.

We are a multidisciplinary group that uses statistical models and develops software to study biology. Members have a background in mathematics, molecular biology, computer science, and statistics. Our approaches are scalable to large datasets, balancing accuracy with efficiency. We have expertise in the False Discovery Rate (FDR) and Linear Mixed-Effects models (LMEs/LMMs). Other topics of interest include modeling of protein sequences and longitudinal data. We are interested in collaborative projects!