Dana Pe’er’s quest for Camelot

While most people become jittery and restless after having a shot of espresso, Dana Pe’er, a biology professor and the principal investigator at Columbia’s Computational Systems Biology Lab, can drink some at midnight and sleep like a rock.

Pe’er and her team have developed a method to predict how organisms respond to drugs based on their genetic information. Her findings indicate that personalized medicine—the notion that patients will one day receive treatments that are based on their genome—could become a reality.

“Personalized medicine can happen,” Pe’er said. “I am envisioning the day that you go to the doctor and have the best course of action.”
While most genomics studies tend to rely on data from DNA, Pe’er’s studies rely more heavily on RNA.

“DNA is the same in each cell, but RNA provides a snapshot of what is actively going on in the cell,” Pe’er explained.

The RNA data reveals which genes are expressed—in other words, which genes are turned on or off. Pe’er hopes that her analysis of gene expression will be “a wake-up call to pharmaceutical companies to be inspired to take RNA measurements.”

Pe’er and her team used yeast as a model organism to test the drug resistance of 94 drugs on different strains of yeast. The team used generally applied human treatments such as fertility, anti-fungal, anti-depressant, and anti-cancer drugs.

First, they determined the drug resistance of 104 strains of yeast based on each strain’s genome sequence (from DNA) and gene expression (from RNA) profile. From these data, Pe’er and her team constructed an algorithm that determines which genes are associated with resistance to each drug.

Pe’er and her team used this algorithm to predict the drug resistance of unexamined strains of yeast by using baseline DNA and RNA data. Bo-Juen Chen, a graduate student in Pe’er’s lab, believes that “the baseline information narrows down the number of genes” related to drug resistance.

The algorithm was able to predict strain resistance for 87 of the 94 drugs tested. To test the accuracy of their predictions, the team genetically modified each strain of yeast by removing the gene that was linked to drug resistance. The team added the drug to the modified yeast and observed that the yeast was no longer resistant.

After a three-year effort, their program—dubbed “Camelot”—was able to pinpoint the gene that caused drug resistance in untested strains of yeast.

Camelot’s goal is to realize the potential of personalized medicine. Genome and gene expression data from an individual could be used to predict the response to a drug before the manifestation of a disease.

Pe’er was recently awarded the prestigious Packard Fellowship for Science and Engineering, which will fund her research to develop a fundamental understanding of how organisms respond to their environment based on their DNA.