Drug Discovery
Computer Screening Uncovers Compounds Against ALS
Neuroscience
Optic Nerve Regrown in Mice
Public Health
Stats Tool Puts
Health Disparities on the Map
Statistical Method Detects Disease Outbreaks
Without Background Population Data
Molecular Modulator of Synaptic Plasticity Revealed
Hot Spots for Genetic Recombination Different in Chimps and Humans
Proceedings of the HMS
Faculty Council
New Appointments to Full Professor
Nominations Invited for Biostatistics Alum Award
Grants Offered in Women’s Health
Honors and Advances
Madras Receives Public Service Award
Students Bag Pharma in Second Year Show
Statistical Method Detects Disease Outbreaks Without Background Population Data
Monitoring a city for disease outbreaks is a tricky business; health workers must balance the expense of investigating false leads against the risk of missing the debut of the next SARS-like infection, while managing a flood of data from hospitals, doctors, and pharmacies. Recent research by Martin Kulldorff may help tilt that balance in favor of early detection by reducing the need for background data on the population being monitored. Kulldorff, an associate professor of ambulatory care and prevention at HMS and Harvard Pilgrim Health Care, published the research on Feb. 15 in the online edition of Public Library of Science Medicine.
Picturing disease outbreaks. Scanning through data from 38 of New York City’s emergency departments, Martin Kulldorf and colleagues identified apparent diarrhea case clusters (not all shown) associated with residential ZIP codes or with hospitals, based only on daily syndromic reports. Data on patients’ chief complaint, home ZIP code, age, and gender were examined from Nov. 15, 2001, to Nov. 14, 2002. Circles represent suspected hospital outbreaks. One residential outbreak includes the entire area of another and nine additional ZIP codes. All signals have p<=0.0027, or one false positive expected per year. (Image adapted by Rachel Meyer)
Kulldorff and colleagues report the development and testing of a statistical process to identify disease outbreaks early, even when very little is known about the population being monitored. “Most methods look at a place and monitor a time series as they occur. We also look at geographic location as a detection variable,” Kulldorff said. The process scans through time, looking for elevated case numbers, and also through space, first examining individual ZIP codes and expanding outward to neighboring ZIP codes within a five-kilometer radius.
Space–time scan statistics like this have previously been used with extensive in-formation about the population at risk, which allows accurate prediction of the expected case number and sensitive detection of outbreaks. But background information is often flawed or unavailable, so Kulldorff wanted to be able to spot outbreaks utilizing case numbers alone.
By taking the daily case data and randomly shuffling the dates and case locations, a large set of data is generated in which no outbreak signal is present, due to the randomization. If the scan detects a signal in the real data that is stronger than in 99.9 percent of these randomized data sets, then health workers can conclude that the outbreak should be investigated, since such a signal would occur by chance only once in a thousand days.
Working with colleagues from the New York City Department of Health and Mental Hygiene, Kulldorff found the technique was able to detect the early onset of influenza season, as well as several small local precursors to wider outbreaks of rotavirus and norovirus. False positives, always a feature of such detection methods, were modest.
Kulldorff explained that his method is not always best. Because the algorithm scans for outbreaks in circles, it has trouble detecting a long narrow outbreak, such as illness caused by a plume of pollution flowing down a river. And because of the way the program searches in time and space, an outbreak has to occur locally first. If people from all over the city become ill simultaneously, a purely temporal detection method (which New York City uses along with the space–time scan) would effectively raise the alarm while Kulldorff’s would not.
Nevertheless, because data on the population at risk is hard to come by, Kulldorff believes the technique can be widely applied. He has made the software implementation, SaTScan, available for free, and even with New York City’s 183 ZIP codes, it runs quickly on common computer hardware. It is now used daily by the Department of Health and Mental Hygiene in parallel with more traditional detection methods.
Molecular Modulator of Synaptic
Plasticity Revealed
A scent, a mathematical formula, an image of a work of art—memories can be built in an instant, but last a lifetime. Memories are made possible by neuronal plasticity, or the ability of a neuron to generate new synaptic contacts; the same phenomenon is required for learning and the proper development of the brain. In the Feb. 17 Neuron, Michael Greenberg and colleagues at Children’s Hospital Boston show that the protein Tiam1, a guanine nucleotide exchange factor (GEF), is involved in regulating synapse development in the mammalian brain. “The finding could help us to better understand neurological disorders and mental retardation—for example, fragile-X syndrome—in which plasticity is aberrant,” said Greenberg, HMS professor of neurology.
In glutamatergic neurons, synapse development and plasticity are mediated, in part, through the N-methyl-D-aspartate (NMDA) type of glutamate receptor. Activation of these receptors evokes both global (i.e., transcriptional) and local (translational) responses within the neuron. In the 1990s, Greenberg and colleagues showed that activation of the small GTPase Ras by NMDA receptors triggers a signal transduction cascade that culminates in transcriptional activation. But what has remained unclear is the identity of the molecular mediator or mediators that evoke local synapse-specific responses.
To address this, research fellows Kimberley Tolias and Jay Bikoff looked for factors that might activate Rac1, a GEF-dependent GTPase known to stimulate remodeling of the dendritic arbor. By promoting actin polymerization, Rac1 stabilizes the cytoskeleton that supports dendritic branches and spines. By turning on Rac1, any GEFs activated by NMDA signaling could modulate plasticity.
Bikoff and Tolias, joint first authors on the Neuron paper, found that of the 60 or so GEFs in the human genome, Tiam1 is the GEF that fits the bill. They found that the protein congregates in hippocampal synapses and binds to NR1, one of the subunits of the NMDA receptor. They also found that stimulation of the NMDA receptor by glutamate leads to phosphorylation of Tiam1 and activation of Rac1.
To confirm that Tiam1 is indeed involved in synapse development, the authors turned to RNA interference to “knock down” levels of specific mRNAs in the cell. RNAi showed that Tiam1 is necessary for the normal development of dendrites, spines, and synapses. Significantly, knock-down experiments also showed that Tiam1 is essential for glutamate-induced dendritic-spine formation and phosphorylation of the kinase Akt, which regulates translation of mRNA. Taken together, the results indicate that Tiam1 plays an important role in local synaptic changes.
Hot Spots for Genetic Recombination Different in Chimps and Humans
Genetic recombination is a key to evolution in sexual species, but only in 2004 did researchers determine that recombination events in humans cluster in pockets instead of being evenly spread across the genome. Now a team led by David Altshuler, an HMS associate professor of genetics at Massachusetts General Hospital, has determined that these recombination hot spots are very different between chimpanzees and humans, despite genetic similarities between the species. Altshuler reports the work with colleagues from HMS, the Broad Institute, and the University of Oxford in the Feb. 10 online Science.
In recombination, DNA strands break, allowing long sections of the genome to swap positions between one parental chromosome and the other. A conversation between Altshuler, who is also the director of the Broad’s Program in Medical and Population Genetics; first author Wendy Winckler, a PhD candidate in the HMS Department of Genetics; and co-author David Reich, HMS assistant professor of genetics, led to an effort to compare chimpanzee and human recombination events. “We figured they would be pretty similar. You’d expect the same pattern of recombination hot spots, probably determined by the sequence,” Altshuler recalled.
In chimpanzee DNA, Winckler found sequences homologous to six known human hot spots, but there was no evidence of similar high recombination rates in the chimps. Twenty-one more hot spots were identified in randomly chosen segments of human or chimp DNA that had been sequenced in multiple individuals. The team found again that these new areas could not be matched across species. The two primates, which share about 99 percent of their DNA sequences, shared almost no recombination hot spots.
Winckler says the results raise many questions about the nature and evolution of recombination. If hot spots are not tied to DNA sequence, researchers will need to examine other mechanisms, such as chromatin structure or nuclear architecture, to determine how the recombination machinery is directed. If hot spots arise and then disappear faster than the underlying DNA sequence changes, what is driving their rapid generation and quenching, and what does it mean for evolutionary change? It seems likely that the research area itself will remain a hot spot until these answers are found.
top