PROJECT SUMMARY AND ABSTRACT
This project endeavors to build a nanosensor array platform technology to detect whole disease fingerprints from
patient biofluids to facilitate diagnosis and biomarker discovery efforts in pancreatic cancer. Pancreatic cancer
is currently the fourth leading cause of cancer-related mortality in the US. The ability to diagnose pancreatic
cancer at early stages would allow many patients to be actively treated, thereby greatly improving their outcomes.
Serum biomarker measurements have been widely used as diagnostic/prognostic indicators, but many markers
are not sufficient for specific assessments of disease states. Major factors limiting precise diagnosis us-
ing these biomarkers include their low sensitivity at high specificity for diseases and the overall dearth of estab-
lished molecular markers. Therefore, innovative approaches to improve disease-state specificity/sensitivity and
biomarker discovery efforts are needed to achieve accurate identification of many conditions. I believe that the
differentiation of disease from normal biofluids may be achieved by the detection of a “disease fingerprint” by
collecting large data sets of molecular binding interactions to a diverse set of moderately selective sensors.
I will build a sensor array comprising organic color centers (covalently-modified carbon nanotubes) stabilized
with DNA to transduce subtle differences in physicochemical properties of molecules in biofluids. With sufficient
diversity, the sensors can differentiate biofluids by disease status with the aid of machine learning pro-
cesses. This platform will also be used to facilitate biomarker discovery efforts. In preliminary data, I discovered
that the responses collected from hundreds of patient samples and interpreted by machine learning algorithms
can beat established serum biomarker measurements. I plan to leverage this technology to develop a robust
diagnostic sensor platform to acquire disease fingerprints of pancreatic cancer in patients biofluids to significantly
increase sensitivity and specificity over single biomarkers and to accelerate biomarker discovery processes. I
propose to investigate: 1) the potential of this technology for the early detection of pancreatic cancer, 2) the
molecular mechanism of the response, and 3) the potential for this platform to enable the discovery of new
biomarkers. In the 2-year mentored (K99) period of the award, I aim to develop a machine perception nanosen-
sor technology with the focusing problem of pancreatic cancer detection and establish the selection rules in the
sensor array construction and the workflow of machine learning-based model development. For the 3-year in-
dependence (R00) period, I aim to systematically investigate how to render the machine learning models trans-
parent to understand the mechanism of high prediction accuracy and discover effective biomarker combinations
for clinical validation studies. Successful completion of the proposed work will result in a validated platform to
enable concomitant identification of early disease states and acceleration of protein biomarker discovery pro-
cesses in pancreatic cancer.
