Once heralded as the holy grail, the capability of obtaining a comprehensive list of genes, proteins or metabolites that are different between disease and normal is routine today. And yet, the holy grail of high-throughput has not delivered so far. Even though such high-throughput comparisons have become relatively easy to perform, understanding the phenomena that cause the disease is as challenging as ever, if not more so. Signaling and metabolic pathways are complex graphs describing genes signaling and biochemical reactions that take place in various subsystems of the organism. The current methods that aim to help us understand the underlying biological phenomena by using the measured differences to identify significantly impacted pathways are rather unsophisticated. Many if not all such methods often treat the pathways as simple sets of genes, and either ignore or under-utilize the very essence of such pathways: the graphs that describe the complex ways in which genes interact with each other. In previous work, we have addressed these shortcoming by developing pathway analysis algorithms that take into account not only the number of differentially expressed (DE) in a pathway, but also the topology of gene interactions captured by these pathways. These algorithms have gained wide acceptance (e.g., SPIA in Bioconductor and Pathway Express) and are the basis for Advaita's recently-introduced Pathway Guide. In the presently proposed work, we will apply and extend this system biology approach to critical problems in personalized medicine and drug discovery. To realize the much-heralded promise of personalized medicine, it is necessary to (a) accurately recognize and distinguish disease subtypes, (b) apply these diagnostic signatures at the level of the individual patient (and tissue type) over time, and (c) identify therapeutic interventions that target them specifically. In this FastTrack proposal we will: i) develop a method and tool able to identify the pathways that are significantly impacted in individual patients/samples (Phase I);ii) develop an analysis technique able to discover and characterize disease subtypes and patient subgroup;iii) develop an analysis technique that can identify pathway signatures for drugs and conditions, iv) develop an analysis technique able to identify new targets for existing drugs;v) test and validate the methods and techniques developed above using real data and feedback from our collaborators. The expected benefits of this project include: the ability to better understand diseases, the ability to identify molecularly distinct subtypes of disease and subgroups of patients and therefore the potential to better tailor treatments and avoid over treatments, and the ability to repurpose existing drugs to new disease indications.
The successful completion of the proposed project will contribute to the public health mission of the NIH in multiple ways spanning from drug development to disease management and diagnosis by leveraging computational strategies of pathway analysis. The communities of users that can benefit from this research include any biological researcher using high-throughput screening methods in any biological domain or organism. Patients can also benefit through the result of better diagnostics, better disease management, and better therapies based on better data.