Xinjun Zhang, Ph.D.

Xinjun (Jun) Zhang, Ph.D.

Assistant Professor of Human Genetics

5912 Buhl Building
1241 E. Catherine St. SPC 5618
Ann Arbor, MI 48109 -5618


Areas of Interest

Admixture, or gene flow between populations, is one of the most ubiquitous and vital evolutionary mechanisms that shaped human evolutionary history and genetic diversity. At the Zhang Lab, we integrate population genetics theories, statistical and computational methods, and empirical human genomics data to solve outstanding questions in human evolution and health that pertain to the interactions between admixture and natural selection. Ancient DNA studies revealed that archaic hominins, such as Neanderthals and Denisovans, admixed with modern human ancestors and facilitated local adaptations in some populations – a phenomenon known as Adaptive Introgression. Most state-of-the-art methods detect adaptive introgression by identifying outliers in one or more summary statistics, which is vulnerable to a high false-negative rate. Non-adaptive processes that can mimic genomic signals of adaptive introgression are also typically unaccounted for in the null models of existing methods, which can inflate false positive rate. Our lab developed a machine learning method called MaLAdapt, which combines information from biologically meaningful features to capture a powerful composite signature of adaptive introgression across the genome. Compared to existing methods, MaLAdapt is especially powerful at detecting adaptive introgression with mild beneficial effects, and is robust to non-adaptive confounders and demographic misspecification. Furthermore, MaLAdapt outperforms existing methods based on validation of simulations and empirical signals. We applied MaLAdapt to empirical human genomic data and discovered novel adaptive introgression loci in worldwide non-African populations, including genes enriched in functionally important biological pathways regulating metabolism and immune responses. For more information about research in our lab, please visit

Web Sites