Our
Research
Research at the Rabadan Lab
Our main scientific interests lie in modeling and understanding the dynamics of biological systems through the lens of genomics. We are a very interdisciplinary team of mathematicians, physicists, engineers, biologists, and medical doctors with a common goal of solving pressing medical problems. We are currently focusing our work on:
Cancer. Genomic technologies provide an extraordinary opportunity to identify mutations that contribute to the development of tumors. We are mapping the evolution of cancers and uncovering the mechanisms of response or lack of response to multiple therapies. We work with clinicians and experimentalists all around the world.
Infectious diseases. Evolution is a dynamic process that shapes genomes. Our team at Columbia is developing algorithms to analyze genomic data, with a view to understanding the molecular biology, population genetics, phylogeny, and epidemiology of viruses. We are interested in the emergence of infectious diseases, pandemics and uncovering the mechanisms of adaptation of viruses to humans.
Electronic Health Records. Clinical databases constitute a rich and complex source of raw data. We are using the power of statistics and computers to tease out important clinical patterns in these diverse, important datasets. Combining molecular and clinical data illuminates some of the mechanisms underlying complex diseases.
In particular, we develop mathematical, statistical, and computational approaches, which cover the analysis of high throughput data right through to the altogether more abstract identification of global patterns in evolutionary processes. Learn more about the Rabadan Lab and the three main global questions that we are addressing.
Research Projects
Global Patterns Of Recombination Across Human Viruses
Viral recombination is a major evolutionary mechanism driving adaptation processes, such as the ability of host-switching. Understanding global patterns of recombination could help to identify underlying mechanisms and to evaluate the potential risks of rapid adaptation. Conventional approaches (e.g., those based on linkage disequilibrium) are computationally demanding or even intractable when sequence alignments include hundreds of sequences, common in viral data sets. We present a comprehensive analysis of recombination across 30 genomic alignments from viruses infecting humans. In order to scale the analysis and avoid the computational limitations of conventional approaches, we apply newly developed topological data analysis methods able to infer recombination rates for large data sets. We show that viruses, such as ZEBOV and MARV, consistently displayed low levels of recombination, whereas high levels of recombination were observed in Sarbecoviruses, HBV, HEV, Rhinovirus A, and HIV. We observe that recombination is more common in positive single-stranded RNA viruses than in negatively single-stranded RNA ones. Interestingly, the comparison across multiple viruses suggests an inverse correlation between genome length and recombination rate. Positional analyses of recombination breakpoints along viral genomes, combined with our approach, detected at least 39 nonuniform patterns of recombination (i.e., cold or hotspots) in 18 viral groups. Among these, noteworthy hotspots are found in MERS-CoV and Sarbecoviruses (at spike, Nucleocapsid and ORF8). In summary, we have developed a fast pipeline to measure recombination that, combined with other approaches, has allowed us to find both common and lineage-specific patterns of recombination among viruses with potential relevance in viral adaptation.