Cancer metabolism is a complex network of perturbations to essential chemical and enzymatic reactions; however, the past century has seen a largely reductionist approach to understanding this system. While previously this approach was necessary due to technological limitations, current computer age technological advances allow us to survey, model, and explore the biological details of individual cells and populations of cells. Scientific fields, such as RNA biology and metabolism, have experienced massive strides in recent decades with the advent of RNA-seq and mass spectrometry-based metabolomics, yet our ability to contextualize and extract the full extent of these enormous datasets continues to lag and often results in focusing on only a handful of entities from a dataset. This effectively causes ?big data? to become ?little data?. This is problematic as these experiments are often expensive and time-consuming to produce, yet we only use a fraction of the total data produced by a given experiment. For the F99 phase of my proposal, I will address these limitations by leading the development of Metaboverse, a multi-omic computational analysis framework built upon our previous work to contextualize -omics datasets within customizable and global metabolic network representations. This framework will lay the foundation allowing for the exploration of complex forms of metabolic regulation in cancer. For example, we will analyze the ability of metabolic networks to undergo dispersed and low-magnitude regulation, where, rather than one or two components acting as the core regulatory actors, regulation is performed by dispersed groups of genes, proteins, or metabolites. This framework and related regulatory research will revolutionize our ability to more holistically understand temporal metabolic shifts and gene-metabolite intra-cooperativity, as well as ensure we obtain the maximum amount of information from our data. For the K00 phase of my proposal, I will work with a postdoctoral mentor at an NCI-Designated Cancer Center or affiliated institution that will supplement my training in machine learning and network biology to develop models that improve our ability to predict metabolic state from transcriptomic state. Doing so will allow us to harness the vast transcriptomics databases in cancer biology to better understand the role of metabolism across heterogeneous tumor cell populations. My ultimate goal is to become a tenured professor and run an independent, NIH-funded research lab that focuses on computational cancer metabolism research and that develops methods for interrogating this emerging domain of biology.
While cancer metabolism is a robust and well-developed research field, approaches to its holistic understanding are still under-developed and hinder our ability to contextualize these complex metabolic states and their consequences. During the F99 phase, I will develop tools and methods that allow researchers to explore data in a more holistic manner than previously possible, which will be essential to elucidating more complicated regulatory mechanisms within cancer metabolism. During the K00 phase, I will develop novel machine learning algorithms that will improve our ability to predict the metabolic state from the transcriptional state, allowing us to harness the rich transcription datasets found in cancer biology for therapeutic benefit.
