Cambridge Rindge and Latin School
Cambridge, MA

Zandra Ho and Andrew Jones
Cancer Data Science 

Hundreds of cancer cell lines are being screened to understand how cancers respond to different drugs or genetic alterations. While we have gained biological insights from the large-scale Achilles (CRISPR-Cas9 gene knockout) and PRISM (drug application) efforts, we have seldom analyzed the datasets in combination. Hoping to find links between drugs and the genes that they target, Kailash used a statistical technique called canonical correlation analysis to uncover shared patterns between the two datasets. Kailash showed that his results were biologically meaningful by mapping these patterns onto biological pathways with a tool called gene set enrichment analysis. In particular, Kailash found a strong association between a group of drugs and genes in the ERBB pathway, suggesting that this tool could be useful in determining which genes a particular drug targets.