First Place: Using Microscopy to Study Chiral Molecular Interactions on the Nanometer-Scale
April D. Jewell, Graduate Student, Chemistry, School of Arts and Sciences
Allister F. McGuire, Undergraduate Student, Chemistry, School of Arts & Sciences
Charles H. Sykes, Associate Professor, Chemistry, School of Arts and Sciences
All humans are left-handed. Not in terms of which hand we use to hold a pen or throw a ball, but on the basic molecular level. All natural amino acids, the basic building blocks of proteins and DNA, have their atoms arranged in such a way that the compounds are asymmetric; in other words, the molecules have two non-superimposable mirror image forms and are what scientists call “chiral”. Chemists describe chirality using a certain nomenclature; right-handed molecules are R-type and left-handed molecules are S-type. The study of chiral molecules is prominent in many branches of chemistry, as well as other sciences. However, there is a lack of fundamental understanding of the forces that drive the interactions between these molecules. In the Sykes group we use scanning tunneling microscopy (STM), which uses electrons to image individual molecules, to study chiral interactions on metal surfaces, including molecular assembly. By stringing together a series of consecutive STM images, we compose movies showing molecular motion on a surface in real time. This movie portrays a cluster of asymmetric molecules, butyl methyl sulfide, moving on a gold surface. The left image shows the real data, falsely colored, while the right image is a schematic aid used to help interpret the data.
At the start of the movie, the two R molecules on the left are static, while the S molecule on the right is spinning. The movie was acquired at 80 K (-193 oC); even so, there is sufficient thermal energy for these molecules to move on the surface, interacting with the surface and neighboring molecules. This STM movie shows the spinning S molecule approach the static R molecules, which subsequently invert their chiralities to become S molecules. They do this by literally flipping over on the surface, akin to flipping a mattress on a bed frame. At the end of the sequence, the three molecules are nested, and all are of S-type chirality. These data clearly shows that the nesting between R-R or S-S molecules is favorable over R-S nesting and give important insight into the fundamental chiral interactions taking place in this system.
Second Place: GDA – Visualizing a Hierarchical Structure of Disease Terminologies
December 2011-February 2012
Jisoo Park, Doctoral Candidate, Computer Science, School of Engineering
Keith Noto, Postdoctoral Researcher, Computer Science, School of Engineering
Heather Wick, Research Coordinator, Computer Science, School of Engineering
Donna Slonim, Associate Professor, Computer Science, School of Engineering
The development of fetal organs and systems plays an important role in later human health. We are interested in identifying connections between sets of genes related to developmental processes and those implicated in disease. To find these connections, we rely on the Medical Subject Headings (MeSH) tree, a hierarchical taxonomy of disease categories. We developed a new method of identifying diseases or disease groups whose associated genes are enriched for a pre-defined gene set. We applied this method to eight sets of genes related to human developmental processes. The gene sets are derived from the Gene Ontology (www.geneontology.org) and our DFLAT project (http://dflat.cs.tufts.edu). We assess the significance of identified overlaps via permutation.
This analysis created a large data set that required visualization tools for adequate exploration. The key challenge in visualizing this data is that there is important information available at multiple scales. For example, a high-level analysis shows that cardiac development genes are more likely than brain development genes to be linked to cardiac diseases. Close-up inspection, however, shows a significant connection between heart development and genes associated with alcoholism. Drilling down further to assess the function of the shared genes points to specific neurological signaling pathways, and might lead the researcher to potential mechanisms of cardiac complications in alcoholism or of heart development anomalies in fetal alcohol syndrome.
The advantage of our visualization is that it enables rapid identification of gene-disease associations at different scales. It starts by providing users with abstract views of the tree structure using simplified triangular forms rather than complex rendering technologies. Although triangles do not represent the full details of a tree structure, the simplification of trees to triangles allows users to easily identify the general nature of disease associations with specific gene sets at a glance. Clicking on a specific triangle (representing a broad class of related diseases) brings users to a new page that shows a detailed tree structure. This tree includes all the significantly enriched diseases in the class and the supporting data. With this tool, we hope to empower researchers to discover hidden relationships between developmental processes and diseases.
Third Place: Food Assistance and the Economy: A Data Visualization
September 2011-February 2012
Parke Wilde, Associate Professor, Friedman School of Nutrition Science and Policy
Jonelle Lonergan, Distance Learning Coordinator, Friedman School of Nutrition Science and Policy
The Supplemental Nutrition Assistance Program (SNAP), formerly known as the Food Stamp Program is the nation’s largest food assistance program and an important part of the social safety net more generally. Because SNAP is a mandatory or entitlement federal program, the caseload and program cost can fluctuate from year to year in response to economic conditions, sometimes surprising and even distressing legislators and administrators. Caseload trends are closely scrutinized, and there is high-level interest in understanding factors that influence program participation.
Research has shown that economic conditions and policy changes are both important determinants of SNAP participation levels. Researchers use state-level data to investigate how the caseload responds to economic and policy variables that change over time and differ across states. Traditional graphical illustrations of such data are either time series or cross-sectional:
- Time series illustrations. When the unemployment rate rises during recessions, workers lose earnings and more families become eligible for SNAP benefits. During economic recovery, the SNAP caseload falls back down again.
- Cross-sectional illustrations. At a single point in time, states with higher unemployment rates tend to have a higher percentage of the population participating in SNAP.
This data visualization improves on these traditional static illustrations by showing SNAP participation dynamics, both over time and across states. The data come from USDA’s Food and Nutrition Service, and the visualization engine was developed by Hans Rosling and released as a Google gadget. The online dynamic bubble plot allows the reader to put the data into motion. The reader may select states of interest and press “play” to view the ebb and flow of SNAP participation during recessions and economic expansions. This approach illuminates the real data underlying econometric analyses of caseload changes, it better illustrates how economic conditions influence the caseload, and it shows the diversity of economic conditions and food assistance participation outcomes across the United States.