Mosquitoes, Worms and You
Human health cannot advance if we only study humans. Consider the work of George Dimopoulos and Alan Scott. Dimopoulos, a molecular entomologist, is analyzing the sequence of the Aedes aegypti mosquito, which transmits dengue, yellow fever and other diseases. Depositing synthesized mosquito DNA onto a glass slide called a microarray, Dimopoulos sees which segments bind to the mosquito's RNA. When one does, it's a real gene. His team has confirmed 9,143 of the mosquito's 15,000+ genes—an essential step for finding Aedes' weak points. Similarly, molecular parasitologist Alan Scott uses sequencing and microarrays to characterize a filarial parasite's genome. (Filarial worms infect more than 120 million people and cause elephantiasis.) Besides developing a vaccine, Scott hopes to learn how the worms survive in the body's lymphatic system, home to our immune response cells. This insight could help prevent autoimmune diseases and reduce the body's rejection of transplanted organs, says Scott.
Taming Information Overload
A decade ago, researchers measured gene expression one gene at a time. Now they use microarrays to measure expression for tens of thousands of genes simultaneously. And the technology is also used to genotype up to a million SNPs at once, identify transcription binding sites, detect methylation and so on. Typical experiments result in hundreds of thousands of numbers per sample. But with the surge of data came more noise and less signal. That's why scientists turn to biostatisticians like Rafael Irizarry to help sift the mountains of digital data for accurate results. A bent glass slide, an errant thumbprint or other contamination can ruin an experiment. Irizarry uses "exploratory data analysis" to discover problems in the data and then applies different models to extract the most useful information. "It turns out you need to be imaginative and creative to do it right," says Irizarry.
