Alejandro Ochoa

This paper introduces CorSym, an unbiased estimator for Pearson correlation in exchangeable variable pairs, demonstrating its effectiveness in reducing order bias in assortative mating studies.

This paper shows that meta-analysis of GWAS can be biased by cryptic relatedness between studies, especially in family-based datasets, and recommends avoiding such analyses.

This study shows that biased kinship estimation in structured populations leads to heritability estimation bias, with Popkin being the only unbiased estimator.

This study compares PCA and LMM in genetic association models, finding LMM outperforms PCA, especially in handling complex relatedness and environmental effects in multiethnic human data.

This study shows genetic association models like PCA and LMM are robust to common kinship estimation biases in structured populations.

This paper improves estimation of FST and kinship under complex population structures by introducing a new framework that addresses biases in existing method-of-moments estimators.

The paper identifies technical biases in STARR-seq data, develops a statistical model to correct them, and improves detection of regulatory elements with better precision and recall.

This study compares PCA and LMM for adjusting relatedness in genetic association studies, finding PCA performs well with large samples and no close relatives.

This paper introduces a generalized FST definition for arbitrary population structures, extending it to account for arbitrary probabilistic dependence among individuals.

The paper introduces new kinship and FST estimators, showing higher genetic differentiation in global human populations compared to existing methods.

The paper introduces a probabilistic method for domain prediction that considers directional domain context, improving prediction accuracy by 15% over existing methods.