Model Systems: Disease Diagnostics
Leroy Hood
is developing the technical and computational tools that will be foundational to blood diagnostics through use of prostate cancer and other organ perturbations (disease and environmental) as a models.
Significance:
We have obtained proof of principle of the dynamics of organ-specific blood molecular fingerprints for liver toxicity as well as in the brain with prion disease. We have developed better proteomics technologies for which we are now developing a microfluidics platform and the isotopic labeling of poly-tryptic proteins.
We have successfully used relative expression data to develop a classifier that differentiate GIST and LMS at the 99% level with just two transcript markers. We believe this approach will be invaluable in distinguishing closely related cancers and in reducing the complexity of the number of markers required to make a particular diagnosis. In addition we are developing algorithms to decipher dynamic blood molecular fingerprints using Halobacterium as a model.
Research and Results:
We have pursued several different lines of investigation to discover organ-specific protein biomarkers or to improve the technologies for finding them. First, we are exploring the powerful new MRM-mass spectrometry approach using isotopically labeled peptides identified from organ-specific transcripts against the transcriptomes of about 40 other organs, both for human and mouse, to determine which transcripts are primarily expressed in the liver. We have approached this problem by significantly improving a major approach to reducing by orders of magnitude the complexity of blood proteins, the glycocapture technique of Aebersold, to the point it is 10-100 times more effective at obtaining N-linked tryptic peptides (see Sun, et.al, 2006). We have also explored a powerful new approach to obtaining the isotope labeled peptides. Mass spectrometry work is being done in collaboration with the Proteomics Core.
Second, we are collaborating with Lumera to experiment with a prototype SPR imager 800-element array for detecting protein binding interactions - that is 800 antibody/antigen interactions can be measured simultaneously. We have demonstrated the power of organ-specific blood proteins to record toxin-perturbed network changes.
A third area of research is our collaboration with Dr. Jim Heath at Caltech to conduct high-throughput typing of the organ-specific transcripts in the blood proteome using DEAL, DNA-labeled antibody libraries. These chips have the potential to be used to measure liver-specific proteins that change in response to toxin injection.
Finally, we developed or are developing two powerful new computational tools for determining the quantitative changes in blood protein concentrations during disease or toxin exposure. . In the first case we have developed a powerful new classification technique using transcriptome data to identify just two markers that have almost 99% ability to distinguish between two histologically very similar tumors and in the second we are using the global data sets of Halobacterium as a test bed to determine whether the changes in these secreted protein concentrations with environmental or genetic perturbations can be used to identify the nature of the perturbed networks.
Plans:
We plan to continue to develop the technologies such as blood proteomics, SPR, and MRM-mass spectrometry, and DEAL technology mentioned above. We will continue our efforts to develop software for reducing the number of markers necessary to make distinguishing diagnoses and we will continue to use Halobacterium as a model for learning how to read the dynamically changing behavior of disease-perturbed networks.
|
Institute for Systems Biology
CSB Home
Center for Systems Biology at the Institute for Systems Biology
