ABSTRACT

Cells receive and respond to signals in their environment, and these signals are often disrupted in disease. Experiments can help understand how proteins interact with each other to alter the cell's behavior; however deciding which proteins to test in an unbiased manner is challenging. Networks, or graphs, are commonly used to represent interactions among proteins, where proteins (nodes) are linked by pairwise interactions (edges). While network-based methods have been popular for many years, predictions from these methods are often challenging to interpret and the tools have not been made easily accessible to biologists, dramatically slowing the potential pace of scientific discovery. The goal of this research is to develop novel methods that more closely reflect the biological questions posed by experimental biologists, and enable the adoption of such tools by the scientific community. This work will be accomplished at a primarily undergraduate institution (PUI), and students who wish to pursue careers in biology need computational training. The project will establish a program for undergraduate training in computational biology at PUIs through local and national initiatives that support both student and faculty development. This project will offer frameworks for (a) introducing computational biology to undergraduates through conference attendance and (b) implementing computational biology activities and courses for undergraduate biology programs with limited resources. Results from this project can be found at http://www.reed.edu/biology/ritz/research.html.


Cells respond to their environment using a series of protein-protein interactions, collectively referred to as signaling pathways, that transfer extracellular signals to the regulation of target genes. Computational methods that describe signaling pathways as graphs have been critical hypothesis-generation tools for understanding the relationship among proteins in cellular signaling response. This project identifies a unifying concept in graph theory -- that of computing directed, connected paths in graphs -- and applies this idea to signaling pathway analysis questions posed in multiple fields of biology. Novel path-finding algorithms will be developed to generate mechanistic hypotheses of active signaling, using dysregulated signaling in disease as a case study. These path-finding algorithms will be applied to signaling pathway analysis in cellular and developmental biology, including pathways that regulate changes in cell shape (morphogenesis) and eye development (retinal neurogenesis). Close collaborations with biologists will help inform the development of easy-to-use tools and broaden their applicability to other fields. The final aim will establish hypergraphs, a generalization of directed graphs, as an improved mathematical representation of signaling. The collection of novel methods produced by this project, along with a demonstration that these tools serve as hypothesis generation engines for other fields in biology, will be a significant step towards accelerating the hypothesis generation-validation-testing research cycle. Biological contributions by adopters of these methods will exponentiate this work's impact on scientific knowledge and discovery beyond the computational contributions in this project.

This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.

Please report errors in award information by writing to: awardsearch@nsf.gov.