Khiabanian Lab

introduction | research | team | publications | software | lab news, etc. | contact

Our goal is to identify markers of selection and progression in cancer, especially through studying the dynamics of tumor heterogeneity, as a model of clonal evolution. Clonal expansion is a process in which a single organism reproduces asexually to give rise to a diversifying population. It is pervasive in nature, from emerging infectious disease outbreaks, to intra-host pathogen evolution, to unregulated cell growth in cancer. The study of clonal expansions began in the 1940s, when Salvador Luria and Max Delbrück designed a simple system of single-cell organisms to investigate patterns of mutation accumulation. Their rigorous quantitative methodology led to discovering that mutations arise randomly and their numbers follow a distinct probability distribution. We now know that a clonal population diversifies as it expands, enabling it to explore the fitness landscape. Studying the dynamics of genomic heterogeneity can yield insight into when an expansion started, how fast a population evolved, and if specific genomic alterations are selected for in a particular host or treatment regime. Cancer follows clonal Darwinian evolution. As genetic alterations accumulate, fitter clones dominate, ultimately leading to macroscopic disease. The evolutionary behavior of cancer progression can be described through linear or branched models of growth. In linear evolution, genetic lesions of the dominant clone from an earlier phase are present in later phases. In contrast, dominant clones in branched evolution share only partial genetic alterations in different phases. Selective pressures can spur tumor evolution and change the mode of its progression from linear to branched; this may lead to more aggressive and treatment-refractory disease. It is, therefore, imperative to capture the extent of genomic diversity in the sub-population structure. Our previous work elucidated the role of mutated subclones as strong predictors of survival and therapeutic response in pediatric and adults leukemia. More recently, we reported on pervasive hypermutation of regulatory regions as a new layer of genetic alterations that dysregulate gene expression in B-cell lymphoma. We are now focused on developing methods that reveal tumor mutational landscapes that correspond to transcriptional heterogeneity and quantify clonal remodeling during disease development and under treatment. We develop new information-theoretic methods for capturing global structural properties of high dimensional datasets and work on novel biclustering algorithms with underlying statistical metrics that allow inference and classification of phenotype-genotype relationships. We also design bioinformatics approaches that address the challenges in interpreting clinical sequencing data and help resolve subclonal tumor alterations from those originating from the non-tumor component in the microenvironment. Our recent analysis of deep clinical sequencing data from patients with solid tumors has showed that some detected mutations arise from infiltrating hematopoietic cells. These mutations are due to an age-related condition known as clonal hematopoiesis of indeterminate potential or CHIP. Our results have raised the hypothesis that CHIP exhibits a distinct genomic landscape when enriched in tumor microenvironment, evolves under solid tumor treatment, and is correlated with the development of therapy-related adverse sequelae. Through developing and integrating novel computational and experimental methods, we are aiming to demonstrate the significance of molecularly defined clonal analysis of hematopoietic populations as a fundamental predictor of disease transformation and therapy-induced complications.   Motivations … The history of science is marked by the efforts of those who strove for precise observations and aimed to decipher the world with the quantitative language of geometry and algebra. In describing the Solar System, it took more than fifteen centuries of staring at the sky, from Ptolemy to Al-Biruni to Brahe, and eventually to Kepler, to find the best fit to the data. Half a century passed and new tools —the telescope and calculus— were devised before Newton could finally describe a graceful approximation to the fundamental governing laws of celestial bodies. More accurate laws of gravity, and physics in general, would not be discovered for another two hundred and fifty years. The study of astronomy was marked by these important paradigm shifts; the field of genomics is undergoing similar revolutions. With the advent of genomic sequencing and high-throughput methods, we are moving into an era characterized by vast amounts of unbiased data. We can now truly observe Darwin's natural selection, “daily and hourly scrutinizing, throughout the world, every variation, even the slightest … silently and insensibly working, whenever and wherever opportunity offers.” Using quantitative and statistical methodologies, not unlike those used by Gregor Mendel, we are on a path to uncover evolutionary insights that will guide predictive models for risk and effective therapeutic approaches for treatment. As in any time of discovery, we must keep in mind that there are golden lessons to be learned, and that one must feel confused —and sometimes stupid— in order to explain the world. In the words of Ptolemy, “when searching out the massed wheeling circles of the stars, our feet will no longer touch the Earth but, side by side with Zeus himself, we will take our fill of ambrosia, the food of the gods.“   Felix, qui potuit rerum cognoscere causas Atque metus omnes, et inexorabile fatum Subjecit pedibus, strepitumque Acherontis avari. That man is blessed who has learned the causes of things, And therefore under his feet subjugates fear And the decrees of unrelenting fate And the noise of Acheron's insatiable waters. —Virgil