Assistant Scientist, Brigham and Women's Hospital
Recommended by John Quackenbush, Dana-Farber Cancer Institute
NEW YORK (GenomeWeb) – Kimberly Glass started out as a physicist, trying to decide whether to pursue experimental physics or theoretical physics. Both had their appeal: There was the math of theoretical physics, but experimental physics had more direct applications.
Rather than choosing between the two sides of physics, she instead turned to computational biology, which includes a fair bit of math, but also can have direct applications. And so, while working toward her PhD in physics at the University of Maryland, Glass also began working with biologists at the National Cancer Institute, bringing together math and biological applications
"It actually integrated pretty well," Glass told GenomeWeb. "My interest was in complex systems and non-linear dynamics in physics, and biological systems are actually very complex and very non-linear, so it actually worked out pretty well."
These days, Glass is working on integrating different types of biological data — transcriptomic, epigenetic, and proteomic data — together to make coherent gene network regulatory models.
This, she said, was borne of a desire to merge the data like that from the ChIP-chip studies she'd been working on at NCI with what she'd learned about networks in her physics program.
The challenge, she said, is to develop models that are not only right, but also useful. One particular problem is dealing with false positives — with such large volumes of data, something is bound to crop up just by chance.
"What kind of keeps me up at night is: I do something, I get all excited, and I go wait a second [and think], 'What if I change this? Does that just completely destroy everything?'" she added.
For instance, Glass noted that differences have been found between men and women in Alzheimer's disease. Women, notably, get the disease at higher rates than men, but women also live longer. The trick then becomes disentangling whether the difference is important for the disease.
Paper of note
In 2013, Glass was the first author on a PLOS One paper describing a method she developed called Passing Attributes Between Networks for Data Assimilation, or PANDA. It incorporates data on protein-protein interactions, gene expression, and sequence motifs to model how information flows through gene regulatory networks.
It's based on a communications theory, she said, that posits that for communication to occur, there has to be a transmitter to transmit the information as well as a receiver to actually receive that information. In cells, that would meant that not only would a transcription factor have to be active, the site where it binds has to be available — it has to be in open chromatin, for instance — to receive that signal from the transcription factor.
PANDA, then, can be used to reconstruct regulator networks based on protein-protein interaction or transcription factor binding data.
Since the method came out, Glass noted that she's been able to apply it in a number of human systems, including studies of chronic obstructive pulmonary disease, ovarian cancer, and sleep deprivation.
"I'm quite proud of that little piece of work, that algorithm. It came out and was applied in yeast and shown to be useful in yeast, but actually what is much more exciting to me is that I've now applied it in a number of different human systems," she said.
Looking ahead
Recently, Glass has also begun to build network models for individual patients, something she said may become more widespread in the coming years as it becomes cheaper and easier to sequence patients' genomes and develop other omic profiles.
"No longer will you just have these three little biomarkers that predict disease," she said. "You'll actually be able to build these complex networks or models and try to understand the mechanisms behind what's causing that in very personalized way."
And the Nobel goes to…
If she were to win the Nobel Prize, Glass would like it to be for developing a method that can be applied to a variety of research questions.
"I'm really driven by wanting to do something that I love that is fun, which is the math and the modeling, but also having an impact on society," she said. "Being able to actually merge those dreams together… would be really amazing."
This is the ninth in a series of Young Investigator Profiles for 2015 that will appear on GenomeWeb over the next few months.