An algorithm developed in our lab called ARACNe uses gene expression data to identify transcriptional interactions. 

An algorithm developed in our lab called ARACNe uses gene expression data to identify transcriptional interactions. 

Our laboratory studies biology from the perspective of the complex networks of gene and protein interactions that define and regulate cell physiology. Combining computational and experimental techniques, we develop models of these networks to predict regulatory checkpoints — also called master regulators — that are essential to the molecular programs that generate specific cellular phenotypes. We have repeatedly demonstrated that master regulators can serve as effective molecular biomarkers of specific phenotypes and constitute extremely promising targets for selectively eliminating diseased cells. In several cases, these findings are now being investigated in clinical trials.

Using this systems biology framework, our goals are threefold: 1) to advance understanding of the general principles that organize biology at a systems level, 2) to develop and disseminate tools and methods for investigating biology in this way, and 3) to identify more effective strategies for disease prevention, diagnosis, and treatment.

Our specific interests include (click to read more):

We see our perspectives as being complementary to those offered by conventional genomics, as they provide methods for elucidating the mechanistic context in which genetic alterations produce specific cellular phenotypes. Often, we have found, master regulators of biological networks do not necessarily harbor mutations of their own, but coalesce upstream alterations in the genetic landscape of the cell in consistent and predictable ways. Characterizing master regulators thus simplifies the problem of accounting for the myriad possible genetic origins of complex diseases, and reveals functional bottlenecks that constitute a distinct kind of “Achilles heel” of disease phenotypes.

Such functional context can also reveal synergistic relationships between multiple genes and proteins that only when acting together become essential in regulating disease-driving networks. This presents opportunities to develop a rational approach to identifying combination therapies based on cell type and network context.

Consistently, our network-based algorithms have grown into mature methodologies that can address the complexity of human disease regulation to a high level of analytical sophistication. The hypotheses that our methods generate have been experimentally validated at a very high success rate — substantially and consistently exceeding 50%, both in vitro and in vivo.

Our work has revealed, for example, actionable new biomarkers and therapeutic targets for the management of many types of cancer, including glioma, leukemia/lymphoma, and prostate cancer. Our approaches have also delivered unique insights into the regulatory networks that define neurodegenerative diseases such as Alzheimer’s and ALS, and the programs that control the differentiation of stem cells along different lineages. Recently, we have also begun investigating the potential effectiveness of using systems biology approaches within a clinical context for cancer diagnosis and, possibly, treatment. Our results are indicating that considering the organization of regulatory networks in individual patients’ tumors could help to identify more effective targeted therapies and combination therapies, and ultimately improve precision medicine.

In addition to developing novel computational algorithms, we conduct extensive wet lab experimentation to evaluate our algorithms rigorously, validate their predictions, and test their relevance to living systems. We also collaborate closely with the JP Sulzberger Columbia Genome Center to integrate high-throughput screening and next-generation sequencing technologies into our work, as well as with investigators with expertise in other experimental methods and in biological fields in which our approaches can contribute new insights. 

Modeling Cell Regulatory Networks >