Thesis Defense - Hamady

We live in a world teeming with microbial life. Microbes occupy every known environment on earth, and trillions live on and inside our own bodies. Although microbes comprise the vast majority of biodiversity on the planet, we are only beginning to realize the importance of our microbial partners. Recent advances in sequencing technology (e.g. massively parallel pyrosequencing) and techniques (e.g. environmental sequencing, molecular barcoding) allow us to characterize microbial communities at an unprecedented rate, producing massive amounts of biological sequence data. However, using highthroughput sequencing to explore microbial diversity on a large scale presents daunting challenges. I address two of these challenges in this thesis. The first is the computational challenge. Although computers have steadily grown faster and more powerful, clock speeds are now stalling and more efficient, parallelized algorithms are thus needed to keep up with the rapid influx of data. The second is the human-computer interaction challenge. The vast, complex datasets now available are difficult to quickly identify patterns in and derive biological insight from, so improved visualization tools are needed to enable users to understand the data. I then highlight several use cases in which collaborators have applied these techniques to gain insight into microbial communities on an unprecedented scale, with applications to understanding diversity of microbes on the human gut and hand, to understanding which factors most affect microbial communities in the gut of different mammals and in a wide range of physical environments, and to understanding how microbial communities differ between lean and obese individuals. The tools I developed thus provide a new perspective on the microbial world, and are likely to have a substantial impact on our understanding of the role of microbes in human health and disease, and in a wide range of environmental processes. Although many challenges remain, we are now able to acquire and understand microbial datasets on a scale orders of magnitude greater than was possible when I began this work.