Mirarab lab focuses on computational biology with a specific focus on developing methods that target evolutionary analyses on large-scale datasets. These algorithms infer statistically rigorous estimates of evolutionary histories based on genomic data and use the results of such inferences in downstream applications. The techniques used range from classic algorithms (dynamic programming) to graph theory and statistical inference. More recently, we have started incorporating machine learning techniques that can integrate biological domain knowledge. High accuracy and scalability are the main focal points, with the idea that gains in scalability should not come at the expense of accuracy. While all the algorithms have heoretical underpinnings, much attention is paid to the empirical evaluation of methods under challenging conditions. The lab prides itself on developing many tools that are widely used by biologists (e.g., ASTRAL series) and have paved new directions (e.g., Skmer, DEPP). We strive to make these useful for biologists and often hold tutorials and workshops for providing training in the use of the tools. Biological applications explored by the lab include reconstruction of species trees from gene trees (phylogenomics), the study of biodiversity using low coverage genomic data (genome skimming), metagenomic analyses using phylogenetic and machine learning approaches, HIV transmission network reconstruction, and large-scale multiple sequence alignment. What unites these wide-ranging applications is their reliance on evolutionary trees as an underlying model.